^ ^
TIME-SERIES DATABASE SETUP AUTOMATIC GENERATION
METHOD, SETUP AUTOMATIC GENERATION SYSTEM AND
MONITORING SERVER
Claim of priority
The present application claims priority from
Japanese application serial no. JP2012 - 258119, filed
on November 27, 2012, the content of which is hereby
incorporated by reference into this application.
10 Field of the Invention
The present invention relates to a time-series
database configuration automatic generation method, a
generation system thereof and a monitoring server,
particularly, to a technology for assisting
15 operational management work to an operations manager
of the time-series database.
Background Art
An operations manager of a computer system
20 monitors a fault that influences a service operated on
the system, and abnormality pre - indicating a portent
thereof. Further, when the fault or the abnormality
is detected, the operations manager analyzes causes
thereof, and takes measures when necessary.
25 Software for assisting such a service management
^> work includes monitoring software. The monitoring
software includes software for assisting fault
detection and a cause analysis. First of all, the
monitoring software has a function for communicating
30 with hardware or software of a monitoring target, and
collecting log data showing a system operation
situation. The log data includes numeric data such as
a CPU utilization rate, or character string data such
as a text message presenting an application state,.
35 In recent years, monitoring software for collecting a
variety of the log data to allow search is also
present. Second of all, the monitoring software has a
function for processing the log data into the form of
a line graph or the like, and displaying the processed
40 form on a screen of an operations manager terminal.
A monitoring mechanism has been so far
constructed for every log data or every service.
However, in connection with an influence of system
integration by recent virtualization technology, an
45 integrated monitoring mechanism for unitarily managing
a plurality of services has been required.
One embodiment of monitoring software that
achieves such integrated monitoring includes a
monitoring platform. In the monitoring platform, a
50 function is provided for putting different kinds of
^o
log data of different services into one place and
managing the log data. In the monitoring platform,
upon request from the operations manager, the Log data
stored on a log database are searched, and the results
55 are presented to the operations manager.
In the log database, capability of storing
different formats of data is required depending on
kinds of services or log data. Technology for
flexibly storing such data exists. Patent document 1
60 (JP-A-2011-134342) discloses a method for, targeting a
database preliminarily needing a definition of a
schema (stored data structure), such as a relational
database (hereinafter, RDB), storing schema data
variable to a defined fixed schema. In the
65 technology, first, a plurality of data field columns
for variable data are preliminarily prepared upon
defining the fixed schema. Then, definition
information in different data formats for every data
type is managed outside the schema, thereby allowing
70 storage of values in different data formats to an
identical field column. As a method different from
the method in Patent document 1, a method also exists
for storing in a hash table a key and a value of data
in different formats to store a binary of table data
75 thereof in a single column of a relational database.
^ ^
Furthermore, a new database provided with schema
flexibility as referred to as a document-oriented
database has recently appeared. The document-oriented
database provides a basic function almost similar to
80 the function of RDB, such as storage, search and
indexing. However, in the document-oriented database,
data are not stored in a tabular format as in the RDB,
but are stored in a combination of a key and a value.
Therefore, a preliminary schema definition becomes
85 unnecessary. In the document-oriented database, the
method for storing the key and the value, as is
different from the art disclosed in Patent document 1
for preliminarily defining what data to be entered
into what column, has an advantage of capability of
90 corresponding also to a case where a data column
definition or sequence is changed later.
Meanwhile, as technology for reducing human
resources in configuration, a method exists for
automating a general database configuration. Patent
95 document 2 (JP-A-2003 - 228570) and Patent document 3(JPA-
H6-215037) disclose arts for evaluating a database
search index from search frequency to an attribute to
perform automatic adjustment. Patent document 4(JP-AH7-
85093) discloses an art for evaluating frequency of
100 utilization frequency of a preregistered index
candidate from search frequency to sequentially select
an index from a superordinate. Patent document 5(JPA-
HlO-111819) discloses an art for automatically
generating an index in which a cumulative value of
105 search response time and insertion response time is
minimized.
Summary of the Invention
In order to utilize a log database of a
110 monitoring platform, configuration of the log database
is required to be performed. The configuration
includes a period configuration as to "what log data
are to be stored for how long of a period." Moreover,
when log data search is desirably performed at a high
115 speed, configuration of data structure including an
index for speeding up is required to be performed. An
operations manager by oneself is required to determine
the configuration of the log databases.
However, determination or adjustment of the
120 configuration of the log database has been complicated
and laborious according to a conventional system. The
reason is that the period needing storage is different
depending on kinds of services or log data. Moreover,
the reason is that search conditions utilized during
125 monitoring and response performance required therefor
•
are different for every kind of log data.
Furthermore, quantitative understanding of an effect
and an influence due to configuration changes has been
quite difficult, and therefore confirmation has been
130 difficult as to whether or not the configuration as
requested by the operations manager is made.
Therefore, a resource such as storage has been so
far excessively consumed in the log database in order
to secure performance while leaving data during a
135 sufficient period. Alternatively, excessive data
deletion or performance shortage has been caused. In
order to prevent such an issue as much as possible, an
expert who has monitoring know-how has distinguished
and examined the configuration, human resources have
140 been consumed. As a result, management of the log
database has been cost.
Even if the methods described in Patent documents
1 to 5 are applied to the log database of the
monitoring platform, generation of a suitable
145 configuration is difficult. As a premise of the
reason, the log database has a feature of the
conventional database, and also a feature for
sequentially accumulating data provided with timeseries
information to perform search such data. The
150 feature occasionally causes an inconvenience in the
^B
background art. For example, in Patent documents 1 to
5, a data utilization period being one of
characteristic information of the log database is not
taken into consideration. If a log data storage
155 configuration or an index configuration is
automatically generated without using the data
utilization period, indexing is performed for all
stored period. However, the log data or the index is
actually utilized only by a specific period. Thus,
160 the resource is excessively consumed by unnecessary
index data outside the period. Thus, even if the
configuration is automatically generated using a
publicly known art, reexamination has been required
for the configuration in consideration of the
165 utilization period using human resources. Moreover,
the publicly known examples are insufficient in a
function for the operations manager to confirm the
effect and the influence of the configuration to
assist selection. Therefore, reduction has been quite
170 difficult for human resources taken in configuration
selection work.
One of the problems to be solved according to the
present invention is to provide a log database
configuration automatic generation method for
175 automatically generating configuration changes
^B
proposal in consideration of the data utilization
period being one of characteristic information on the
time-series database in the monitoring platform, and
presenting the configuration changes proposal to the
180 operations manager, thereby allowing assisting of
configuration changes work by the operations manager,
and reduction of cost required for log monitoring by
the operations manager, and the configuration
automatic generation system, and the monitoring
185 server.
Assisting of configuration changes work applying
the present invention allows automatic generation of
configuration changes proposal in consideration of a
data utilization period for every monitoring target to
automatic generate configuration information changes
proposal, and therefore allows reduction of waste of a
resource of a time-series database of a monitoring
platform, and reduction of cost required for
monitoring by an operations manager.
190
195
200
Brief Description of the Drawings
Figure 1 is a drawing showing an example of a
monitoring platform to which a log database according
to the present invention is applied.
Figure 2 is a drawing schematically showing a log
^A
database configuration automatic generation system
related to a first embodiment according to the present
invention.
Figure 3 is a drawing showing an example of log
205 raw data in a first embodiment.
Figure 4 is a diagram showing an example of log
compressed data.
Figure 5 is a diagram showing an example of log
index data.
210 Figure 6 is a diagram showing an example of
tenant information.
Figure 7 is a drawing showing an example of an
operation history.
Figure 8 is a drawing showing an example of
215 current configuration information.
Figure 9 is a drawing showing an example of data
volume information.
Figure 10 is a diagram showing an example of
individual policy information.
220 Figure 11 is a diagram showing an example of
entire policy information.
Figure 12 shows a sequence of configuration
changes proposal generation in a first embodiment.
Figure 13 shows a flowchart of data configuration
225 information generation processing of a data
10
^m
configuration information generating unit in a first
embodiment.
Figure 14 is a diagram showing an example of
configuration information after change in a first
230 embodiment.
Figure 15 shows a flowchart of data configuration
information adjustment processing of a data
configuration information adjusting unit in a first
embodiment.
235 Figure 16 is a diagram showing an example of
difference information of configuration changes.
Figure 17 shows a flowchart of index
configuration information generation processing of an
index configuration information generating unit in a
240 first embodiment.
Figure 18 shows a flowchart of index
configuration information adjustment processing of an
index configuration information adjusting unit in a
first embodiment.
245 Figure 19 is a diagram showing an example of a
configuration changes entire metrics value.
Figure 20 is a diagram showing an example of
configuration changes proposal recommended screen
display in a first embodiment.
250 Figure 21 is a diagram schematically showing a
11
^^
log database configuration automatic generation system
related to a second embodiment according to the
present invention.
Figure 22 is a diagram showing an example of
255 selection information.
Figure 23 shows a configuration changes
readjustment sequence in a second embodiment.
Figure 24 shows a flowchart of policy information
automatic adjustment processing of a selection
260 information analysis unit in a second embodiment.
Figure 25 is a diagram schematically showing a
log database configuration automatic generation system
related to a third embodiment according to the present
invention.
265 Figure 26 is a diagram showing an example of
service information.
Figure 27 is a diagram showing an example of
fault information.
Figure 28 shows configuration changes proposal
270 generation sequence in a third embodiment.
Modes for Carrying out the Invention
First, an outline of a time-series database
configuration automatic generation system according to
275 the present invention is described.
12
^
280
285
290
295
300
The configuration automatic generation system
according to the present invention is achieved by a
computer system as described in embodiments below.
Physical or virtual equipment on the computer system
is to be lent to a plurality of enterprises or
organizations. In the present invention, aggregation
of equipment lent to each enterprise is referred to as
"tenant." In the present invention, monitoring of a
service is performed for every tenant. The equipment
on the computer system in the present invention is
shared among tenants. However, an operations manager
of a certain tenant is not allowed to browse
information of any other tenant to be stored on the
equipment. The reason is that log data for monitoring
the service may include a trade secret of each
enterprise.
Figure 1 is a drawing showing an example of a
monitoring platform to which a time-series database
configuration automatic generation system related to
the present invention is applied. In the monitoring
platform, a function is provided for putting different
kinds of log data of different services into one place
and managing the log data. Moreover, the monitoring
platform shown in Figure 1 is formed as a multi-tenant
type monitoring platform. The multi-tenant type
13
^
monitoring platform refers to a form in which a
monitoring platform provider provides the monitoring
platform with a business operator or business
operators (hereinafter, tenant) as a service (Software
305 as a Service).
Here, an outline of the functional of the
monitoring platform is described using Figure 1.
Monitoring platform 201 held by system 200 of a
monitoring platform provider is connected through
310 network 300 to each of systems 100, 110 and 120 of a
plurality of tenants (A), (B) and (N) using the
monitoring platform. Attention is paid to system 100
of tenant A herein. Tenant A provides a service to
each end user in tenant A by server 101 (A-1 to A-n)
315 on the system 100. Monitoring platform 201 of system
200 of the monitoring platform provider is provided
with time-series database 202 and monitoring server
203. An operations manager of tenant A is accessible
through terminal 102 to server 101 of a monitoring
320 target and system 200 of the monitoring platform
provider. In order to monitor an operation situation
of each service to be provided for the end user in
tenant A, system 200 of the monitoring platform
provider sequentially collects time-series data such
325 as various kinds of log data output or measured on
14
A|
330
335
340
345
350
each server 101 (A-1 to A-n) on monitoring platform
201 to store the data in time-series database 202. In
the present invention, such a database for storing the
time-series data is referred to as a time-series
database. Monitoring server 203 of system 200 of the
monitoring platform provider has configuration
management function 204 for time information of timeseries
data or the like, time-series data utilization
and analysis function 205, and changes proposal
generation and registration function 206 for the time
information or the like for every monitoring target,
and is provided with a database for management and
configuration in which various kinds of information
for management and configuration are held (see Figure
2) .
Time-series data utilization and analysis
function 205 searches time-series data upon request of
monitoring by the operations manager, and outputs
results thereof to terminal 102 of the operations
manager. Changes proposal generation and registration
function 206 searches, utilizes and analyzes, based on
information on the database for management and
configuration, the time-series data stored in timeseries
database 202, and generates, upon request of
change regarding the time-series data from an
15
^ B
operations manager of tenant A, or from a monitoring
platform side itself, time-series data changes
proposal including time information (for example, a
retention), for example, once a month, presents
355 results thereof to terminal 102 of the operations
manager, and if approval is obtained, changes the time
information of the time-series data.
In addition, according the present invention, the
combination (tenant, data type, instance name) forming
360 a unit for generating configuration changes proposal
is defined as "monitoring target."
Thus, the monitoring platform provider
(operations manager) allows suitable updating of
configuration information regarding conservation of
365 time-series data for every monitoring target with
meeting a requirement of the operations manager of the
tenant. Such updating allows reduction of cost
required for log monitoring (in particular, management
of the time-series database) by the operations manager
370 of the monitoring platform. More specifically,
according to the present invention, configuration
changes proposal in consideration of a data
utilization period are automatically generated for
every monitoring target. Metrics information to be
375 served as a judgment source of an effect and an
16
^g influence by the configuration changes proposal is
presented. Such presentation allows reduction of a
man-hour of log database configuration examination
work. Second, in the present invention, configuration
380 changes proposal are generated by analysis a kind and
a period of log data actually utilized from the
operation history. Configuration changes of timeseries
data conservation format is performed according
to the proposal, thereby allowing reduction of
385 unnecessary data on the log database. Moreover,
intensified log monitoring of the monitoring platform
by the operations manager is allowed. According to
the present invention, the configuration changes
proposal in conformity with utilization are generated
390 for every monitoring target. Therefore, suppression
is allowed for loss caused by excessive log data
deletion (fault correspondence man-hour increase,
opportunity loss) that has been so far generated.
In addition, in an embodiment according to the
395 present invention as described below, the log database
is illustrated as the time-series database of the
monitoring platform. However, the embodiment
according to the present invention is not limited to
the log database. The present invention can be
400 applied if the time-series database is accumulated in
17
^1 the order of time series. For example, the present
invention may be applied to time-series database for
accumulating a communication (telephone call) history
(time-series data) in a telecommunications sector.
405 The communication history in which an enormous amount
of data is stored for every user is searched, for
example, for improving service quality. Therefore,
the communication history is similar, in the
utilization, to the log database presented in the
410 embodiment according to the present invention, and
therefore, can be managed by the method of the
embodiment according to the present invention.
Moreover, the present invention can be applied to Web
blog update information, contributed information in an
415 asynchronous short sentence contribution service, a
time-series database for marketing in which timeseries
data such as a buying history using an IC card
is stored.
Hereinafter, a detailed embodiment according to
420 the present invention is described with referring to
drawings.
[Embodiment 1]
A first embodiment of the present invention
refers to an example in which log database
425 configuration changes proposal are automatically
18
^^
generated on occasion of request from an operations
manager to present the configuration changes proposal
to the operations manager.
In Figure 2, a log database configuration
430 automatic generation system related to the first
embodiment of the present invention is schematically
shown. Figure 2 schematically shows a computer system
for achieving the first embodiment.
The computer system is constituted of log
435 database 3, management database 4, configuration
changes proposal database 5, log monitoring server 6,
log configuration management server 8, and operations
manager terminal 7 of a tenant. The relevant
equipment is connected through physical communication
440 line 2 to management network 1.
Log database 3 is used for the operations manager
to manage equipment log data to provide a service. In
log database 3 in the present embodiment, log data are
managed in a predetermined conservation format for
445 every tenant, every kind of log data and every
equipment. Moreover, the log database 3 is to be
operated through operations manager terminal 7 by a
plurality of operations managers of each tenant. The
manager terminal 7 may exist as many as the number of
450 operations managers.
19
^^
Management database 4 stores information for
managing log database 3, and is utilized by each
server and each unit (program) as described later.
Management database 4 holds management information
455 regarding log databases 3, such as tenant information
T400, current configuration information T600 of log
database 3, and data volume information T700 of log
database 3, and a history regarding log database 3,
such as operation history T500 to log database 3.
460 Log monitoring server 6 holds, in a memory
thereof, log collection program 61 for allowing a
server (computer) to function as "log data collection
and storage unit," log monitoring program 63 for
allowing the server (computer) to function as "stored
465 log monitoring unit" and data volume analysis program
62 for allowing the server (computer) to function as
"data volume analysis unit."
"Log data collection and storage unit" (log
collection program) 61 has a function for collecting
470 log data from equipment for service provision. As a
general collection method, the log data are collected
by operating an agent such as Simple Network
Management Protocol (SNMP) on each of equipment, and
allowing communication between the agent and "log data
475 collection and storage unit" 61. "Log data collection
20
^ ^
and storage unit" 61 sequentially registers received
log data into log database 3.
"Stored log monitoring unit" (log monitoring
program) 63 provides a function for searching log data
480 upon request of monitoring by the operations manager
to display results thereof. For example, if "stored
log monitoring unit" 63 receives a monitoring request
from web browser 71 displayed on a screen of
operations manager terminal 7, extracts log data from
485 log database 3 according to a log-data search
conditions included in the monitoring request. Then,
the thus extracted log data are analyzed and processed
according to the monitoring request to display results
thereof on web browser 71. Moreover, "stored log
490 monitoring unit" 63 has a function for recording
through the unit a history of performing a log data
monitoring operation on management database 4.
Hereinafter, the history is referred to simply as
"operation history."
495 "Data volume analysis unit" (data volume analysis
program) 62 has a function for analysis log data
registered into log database 3 to analyze a stored
data volume (hereinafter, referred to as data
conservation amount). An analytical method therefor
500 is described later.
21
A Log configuration management server 8 provides a
function for the operations manager to manage
configuration of log database 3. Log configuration
management server 8 holds, in a memory thereof,
505 configuration management interface 81 and
configuration changes proposal generation program 82.
Configuration changes proposal generation program 82
is a program for allowing the server (computer) to
function as "configuration changes proposal generating
510 unit," and is constituted of five subprograms
including data configuration information generation
program 83 for allowing the server to function as
"data configuration information generating unit," data
configuration information adjustment program 84 for
515 allowing the server to function as "data configuration
information adjusting unit," index configuration
information generation program 85 for allowing the
server to function as "index configuration information
generating unit," index configuration information
520 adjustment program 86 for allowing the server to
function as "index configuration information adjusting
unit" and configuration changes proposal display
program 87 for allowing the server to function as
"configuration changes proposal displaying unit." In
525 addition a category of each unit (each subprogram) is
22
^
530
only one example, and some of the units (subprograms)
may be obviously unified to one, or any one of the
unit (subprogram) may be obviously categorized and
constituted according to each function.
Configuration management interface 81 is a unit
(program) for browsing or changing, upon request of
monitoring from operations manager terminal 7,
configuration of configuration information of log
database 3 of a tenant as managed by the operations
535 manager. During browsing the configuration
information, configuration management interface 81
accesses management database 4, acquires the
configuration information of log database 3 of the
tenant as managed by the operations manager who places
540 request, and displays the configuration information on
a screen. The operations manager places request of
configuration changes by inputting the request of
addition, deletion, overwriting or the like to the
configuration information displayed on the screen.
545 Upon request of the configuration changes from the
operational manager, configuration management
interface 81 accesses management database 4 to update
recorded configuration information. Furthermore,
configuration management interface 81 accesses log
550 database 3 to reflect a new configuration.
23
^^
"Configuration changes proposal generating unit'
82 uses the information stored in management database
4 to analyze a utilization situation of log database 3
of the tenant to automatically generate configuration
555 changes proposal in conformity with the situation.
"Configuration changes proposal generating unit" 82 is
invoked on occasion of request through configuration
management interface 81 from the operations manager to
generate the configuration changes proposal and
560 display the changes proposal on a screen.
In the present embodiment, a means is provided
for automatically generating the configuration changes
proposal of the log database from the operation
history of log monitoring unit 62, the data volume
565 information of log database 3, the current
configuration information or the like. In addition to
the conventional log database, a management database
is prepared in which the history information of the
operation to the log monitoring unit, the data volume
570 information of the log database and the current
configuration information are stored. If the
operations manager of the tenant performs through the
log monitoring unit an operation of log database
monitoring, the log monitoring unit records a history
575 of the operation on management database 4. Moreover,
24
^h
if the operations manager performs through
configuration management interface 81 configuration
changes of the log database, the configuration
management interface records the configuration
580 information on the management database. Furthermore,
if log data of a monitoring target device are
collected on a log database through log acquiring unit
61, log data volume analysis unit 62 reads utilization
information of the data volume of the log database,
585 and records the information on the management
database.
Configuration changes proposal generating unit 82
automatically generates log database configuration
changes proposal based on the information recorded on
590 the management database. In order to perform the
automatic generation, the unit 82 utilizes information
such as the operation history, the information on a
log utilization period and a data conservation format
included in the current configuration information.
595 Moreover, in order to present an effect and influence
of the automatically generated changes proposal to the
operations manager, the unit 82 calculates a metrics
value for evaluating time required for configuration
changes, a metrics value for evaluating performance in
600 connection with the configuration changes, and a
25
^ ^
metrics value for evaluating the data volume in
connection with the configuration changes to presents
the information to the operations manager.
Generation processing of the configuration
605 changes proposal by the function of each of units 83
to 87 (program) is described later in detail.
Hereinafter, the log data stored in log database
3 in the present embodiment are described in detail.
Log database 3 in the present embodiment stores
610 log data for every tenant. In Figure 2, an example is
shown in which log data of a plurality of tenants are
managed on one log database 3.
Moreover, log database 3 in the present
embodiment stores log data in a tabular format.
615 However, the embodiment according to the present
invention is not limited to data in the tabular
format. If data corresponding to each of information
as described later, for example, tree structure data
or text data exist, the present invention can be
620 applied thereto.
An example of the log data stored in log database
3 in the present embodiment includes three kinds of
data conservation formats, including raw data TlOO,
compressed data T200 and index T300. Raw data TlOO
625 are log data per se collected from the equipment for
26
^^
service provision. Moreover, compressed data T200 and
index T300 include information generated by analysis
and processing the raw data TIOO.
In the present embodiment, the information on the
630 tenant, the kind of log data, and the equipment for
service provision is provided for the log data.
Subsequently, a name for uniquely identifying the
kind of log data is referred to as "data type name."
The data type includes data regarding an OS, data
635 regarding middleware (database, Web container or the
like) commonly used from a plurality of applications
on the OS, and data inherent to the application. The
data type name includes "Web event log of service 1,"
"CPU utilization rate of OS" or the like, which is
640 character string data by which the types can be
specified.
Moreover, a name for uniquely identifying the
equipment for service provision is referred to as
"instance name." The instance name might include a
645 host name or an IP address.
In log database 3 in the present embodiment, a
configuration of each conservation format is
determined in a unit of the tenant, the data type name
and the instance name. Hereinafter, a set of the
650 tenant, the data type name and the instance name is
27
^|. referred to as a monitoring target. Data to be stored
in raw data TlOO, compressed data T200 and index T300
are associated with each other by the monitoring
target. In the present embodiment, the tenant, the
655 data type name and the instance name are used as the
monitoring target, but is only an example, and
information different therefrom may be used as the
monitoring target.
Figure 3 shows one example of raw data TlOO.
660 Column T102 includes a tenant name for managing a log.
Column T103 represents time when the log is output.
Column T104 includes a data type name and column T105
includes an instance name. Information presented by
the columns T102, T104 and T105 is monitoring target
665 T120. In an example in the present embodiment, the
information in columns T102 to T105 is essential data
to be provided for all the logs without fail.
Meanwhile, column T106 presents data in which the
content stored is different for every data type name
670 or instance name. Hereinafter, data in which the
content stored is different for every data type name
or instance name are referred to optional data. The
optional data are stored as a pair of a name (key) and
a value. Hereinafter, the key and the value are
675 referred to as an optional data name and an optional
28
•
data value, respectively. Thus, the data are managed
by a pair of the name and the value as described
above, thereby allowing storage of logs in a wide
variety of formats in a table. For example, a first
680 pair of column T106 of row T150 presents that data of
a name "event name" and a value "Actionl" are recorded
on a log. Column T112 presents a byte count per log
data. The byte count includes a total value of byte
counts of various types of essential data and byte
685 counts of optional data of the log, for example.
Here, in Figure 3, in order to simplify an
expression in the Figure, an example is shown in which
raw data are stored in a tabular format. However, the
data may be practically stored in a document-oriented
690 database for storing all the data in a set of the key
and the value. Even in such a case, no significant
difference exists in comparison with a case where all
sets of keys and values are stored in a single column
in the tabular format as shown in the Figure.
695 Figure 4 shows one example of compressed data
T200. Compressed data T200 include data prepared by
compressing raw data TlOO according to a predetermined
conversion procedure. Conservation capacity can be
saved by compressing the data. However, when
700 compressed data T200 are subjected to log data search.
29
•
the data is required to be once decompressed and then
searched, and therefore search needs further time in
comparison with the raw data. Therefore, raw data
TlOO with low utilization frequency or raw data TlOO
705 whose search performance may be late are generally
subjected to a compression target. In the example,
raw data TlOO included within a predetermined period
are collectively compressed, and a mass of the
compressed data is stored as one row (record).
710 Columns T201 to T203 present a tenant, a data type
name and an instance name. The columns correspond to
columns T102 to T104 of raw data TlOO. Columns T204
and T205 present compression start target time and end
target time, respectively. Among raw data TlOO to
715 which columns T102 to T104 correspond, data in which
log measurement time is included between compression
start target time and end target time serve as
compression target data in the record. Column T206
presents a count of raw data subjected to the
720 compression target. Column T207 presents actually
compressed data. Generally, compressed data T200 are
expressed in a binary format, and therefore are
difficult to allow reading. Therefore, the
description of column T207 is omitted. Column T208
725 and column T209 present a byte count before
30
compression and after compression, respectively.
Figure 5 shows one example of index T300. The
index refers to data for accelerating raw data search
on a table. In the present embodiment, a column index
730 for accelerating search format least one data column
in the table is to be used as one example of the
index. Moreover, in target data of the index,
designation is to be allowed for an essential data
name and an optional data name of raw data TIOO. For
735 example, when "log measurement time" being the
essential data name of raw data TIOO is indexed,
search is allowed at a higher speed than usual on
performing designation of specific time or a range of
time to log search conditions. Moreover, when "event
740 name" being the optional data name is indexed, search
at a high speed is allowed on designating the event
name to the log search conditions. A plurality of
data columns may also be indexed. For example, in the
case where "data type name, log measurement time, and
745 an instance name" are indexed in the order, when
refinement is performed in the consecutive order on
search conditions, high-speed search is allowed. When
refinement is performed in the order of the data type
name, the log measurement time, and the instance name,
750 in the order of the data type name and the log
31
measurement time and only in the data type name in the
index, high-speed search is allowed. Meanwhile, in
the case when search is deviated from the order, when
refinement is performed in the order of the instance
755 name and the data type name, high-speed search using
the index is not allowed. Moreover, when a data
column that is not included in the index is added to
search conditions, high-speed refinement is performed
in a column included in the index, and then ordinary
760 refinement is performed in a column that is not
included therein. Columns T301 to T303 present a
tenant, a data type name and an instance name. The
columns correspond to columns T102 to T104 of raw data
TIOO. Column T304 includes information presenting
765 target data of the index. All of rows T350 to T354
are examples of providing column indices for raw data
(rows T151 to T151) in which monitoring targets of log
data includes (tenant A, Web event log of service 1
and aplserverl). However, target data columns are
770 different in rows T350 to T354, respectively. Thus, a
plurality of indices may be configured for one
monitoring target. Column T305 is a column presenting
internal data structure of the index. In column T305,
the data structure of the index per se may be stored,
775 or reference information to data structure may also be
32
Stored. For the data structure of the index in the
present embodiment, use of tree structure referred to
as B+tree that is adopted in many databases is
presumed, and the description on detailed internal
780 structure is omitted. Columns T306 and T307 present a
byte count and a data count of the index data,
respectively.
In Figure 5, all the index data are managed in
one table. However, the embodiment of the present
785 invention is not limited to the system. For example,
a separate index table may be prepared for every
monitoring target.
In the present embodiment, the three data
conservation formats as described above are handled.
790 However, the embodiment according to the present
invention is not limited to the data formats. For
example, results of analysis a plurality of log data
within predetermined time may be stored as cache.
Hereinafter, information to be stored in
795 management database 4 in the present embodiment is
described in detail.
Figure 6 shows one example of tenant information
T400. Tenant information T400 includes data
presenting correspondence relationship between a
800 tenant using log database 3 and an operations manager
33
belonging to the tenant. Column T402 includes a name
of the tenant to whom the operations manager belongs.
Column T403 includes a name of the operations manager.
Figure 7 is a drawing showing an example of
805 operation history T500 relative to log monitoring unit
63. Column T501 includes time and date on which an
operations manager accesses log monitoring unit 63.
Column T502 includes a name of a tenant to whom the
operations manager belongs. Column T503 includes a
810 name of the operations manager who accesses log
monitoring unit 63. Columns T504 to T506 include
information presenting search conditions designated
during browsing. Columns T504 and T505 include a
range of log measurement time designated by the search
815 conditions. Column T504 includes a minimum value of
measurement time, and column T505 includes a maximum
value thereof. In the present embodiment, designation
of a range of measurement time of day on the search
conditions is to be essential. Moreover, log data of
820 the minimum value or more and the maximum value or
less of designated measurement time are to be
searched. Column T506 includes search conditions
other than the log measurement time of day. For
example, row T553 includes conditions when search is
825 performed for log data, in which, within designated
34
measurement time, a data type name agrees with "Web
event log of service 1," an instance name agrees with
"aplserverl," and request processing time is "2000" or
more. Column T507 includes response time in a server
830 required for search under the conditions. Column T508
includes data count hit during search under the
conditions. Column T509 includes an index utilized
for search under the conditions. In the column,
information on a target data column of a utilized
835 index is registered. When any index is not utilized
for search, "unused" is registered. Column 510
presents whether or not any access is made to
compressed data upon the search. When any access is
made to compressed data, "existence" is registered,
840 and when any access is not made thereto, "notexistence"
is registered.
In the present embodiment, the search conditions
as described above are taken as one example, but the
present invention is not limited only to the
845 conditions. For example, the search conditions may
also include designation of sorting conditions.
Figure 8 is a diagram showing an example of
current configuration information T600 of log database
3. In log database 3, a configuration according to
850 the information is reflected. Column T601 includes ID
35
for uniquely identifying a configuration of a log
database. The configuration ID is to be newly
automatically provided upon configuration information
being added. In a processing flow to be described
855 hereinafter, configuration information may be
occasionally added, but the description is omitted for
automatic provision processing of the configuration
ID. Columns T602 to T604 include information
presenting a monitoring target. Column T602 includes
860 a name of a tenant managing log data to be configured.
Column T603 includes a name of a data type of log
data to be configured. Column T604 includes a name of
an instance for outputting log data to be configured.
Column T605 presents a conservation format of data
865 designated for the monitoring target. In the
conservation format, any one of "raw data,"
"compressed data" or "index" is designated. Column
T606 presents a target data column for which an index
is provided. The column is used only when "index" is
870 designated into data conservation format T605. Column
T607 includes data retention to be configured. In the
present embodiment, a range of time starting from
current time is designated in the retention, such as
[from 180 days before to current]. In log database 3,
875 only log data within the range of the retention are
36
880
885
890
895
900
stored, and log data out of range are sequentiallydeleted
.
Here, a raw data configuration, a compressed data
configuration and an index configuration are described
using an example, respectively. Row T650 includes
configuration for conserving raw data in which
information presenting a monitoring target is (tenant
A, Web event log of service 1, and aplserverl) from
180 before to current. Log data outdated from 180
days are, when the compressed data configuration is
made, sequentially compressed, and when the
configuration is not made, are sequentially deleted
from on log database 3. Row T651 includes
configuration for conserving compressed data of a
monitoring target identical with the target of row
T650 from 360 days before to 180 days before. Thus,
an end time point of the retention of compressed data
is configured so as to be continuous with a start time
point of the retention of raw data. Rows T652 to T654
includes an index configuration to raw data having
information presenting a monitoring target identical
with the target of row T650.
Figure 9 is a diagram showing an example of data
volume information T700 of log database 3. In the
data volume information T700, results are stored in
37
^. which log data analysis unit 62 analyzes data of log
database 3 to tabulate a data conservation amount and
count. Column T701 includes ID for uniquely
identifying a log database configuration. Columns
905 T701 to T705 include information presenting a
monitoring target and a data conservation format. The
columns T701 to T705 correspond to columns T601 to
T605 of configuration information T600. Column T706
includes data conservation capacity of log data of the
910 configuration. Results obtained by tabulating the
data conservation capacity stored on log database 3
are stored in the column. Here, a calculation method
for the data conservation capacity is described for
every data conservation format in the present
915 embodiment. When data conservation format T705
includes "raw data," log data analysis unit 62 refers
to raw data TlOO on log database 3, extracts rows in
which information T102, T104 or T105 presenting a
monitoring target agrees with T701, T702 or T703, and
920 counts byte count T112 thereof. When data
conservation format T705 includes "compressed data,"
log data analysis unit 62 refers to compressed data
T200 on log database 3, extracts rows in which
information T201 to T203 presenting a monitoring
925 target agrees with T701, T702 or T703, and counts byte
38
count T209 thereof. When data conservation format
T705 includes "index," log data analysis unit 62
refers to index T300 on log database 3, and acquires
byte count T306 of an index in which a monitoring
930 target agrees with target data columns T301 to T304.
Column T707 includes the data conservation count of
log data of the configuration. The data count is also
tabulated in a manner similar to the operations in the
conservation capacity. However, a row count but not a
935 byte count is counted. Counting is made simply for a
row count in the case of "raw data," compression
target data count T206 in the case of "compressed
data," and data count T307 of the index in the case of
"index," respectively.
940 Figure 10 is a diagram showing an example of
individual policy information T900 of configuration
changes. Individual policy information T900 includes
data that describes a level that each tenant requires
for an individual configuration to be changed upon the
945 configuration changes of log database 3. In the
present embodiment, when configuration changes
proposal generating unit 82 creates configuration
changes proposal, the unit 82 judges whether or not an
individual configuration satisfies the level, and
950 presents only a level that satisfies the level as the
39
configuration changes proposal. In the present
embodiment, at least one of individual policy
information T900 is to be configured for every tenant.
If a plurality of individual policy information T900
955 are configured by one tenant, configuration to be
changed only needs to meet any criterion among the
plurality of individual policy information T900.
Column T901 presents ID for uniquely specifying
individual policy information T900. Column T902
960 includes a name of a tenant who configures the policy.
Columns T903 to T905 include a level of a metrics for
a tenant to require for an individual configuration to
be changed. In the present invention, upon generating
the configuration changes proposal, three kinds of
965 metrics values for evaluating the individual
configuration to be changed are calculated. When the
calculated metrics value satisfies a level presented
in columns T903 to T905, the configuration is adopted.
Hereinafter, the metrics for evaluating the
970 individual configuration to be changed is to be
referred to as an individual metrics.
Column T903 includes a level of individual
metrics Ml for evaluating time required for the
configuration changes, namely, time needed for the
975 configuration changes. The time required for the
40
980
985
configuration changes is an important metrics in
determining the configuration of log database 3. The
reason is that further cost is required during the
configuration changes, in comparison with the normal
time, and a risk of opportunity loss increases, and
therefore the period is desirably shortened as much as
possible. On generation of the configuration changes,
data processing such as batch addition and deletion
processing of log data is performed. During the data
processing, log data unconformity or performance
degradation by lock of data may be generated. In
order to avoid the influence, the configuration
changes are generally performed in a specific time
zone such as a system usage stop day and a holiday.
Whether the processing for the configuration changes
is completed within the time zone becomes a large
metrics for the operations manager. In the present
embodiment, assumed processing time required for the
configuration changes is used as one example of
metrics Ml. In column T903, conditions to be met by a
metrics value used for evaluating metrics Ml are
configured. For example, row T950 presents that a
level of metrics Ml is configured as "less than 60
minutes in time required for the configuration
1000 changes" in a certain individual policy. In the
990
995
41
present case, configuration of a log database
requiring 60 minutes or more is judged to be
configuration that does not meet the policy. In
addition, when setting of a level for the metrics is
1005 not needed, "unlimited" is configured. The
configuration is also similar to the metrics described
hereinafter.
In the present embodiment, metrics Ml is
evaluated using processing time, but is evaluated by a
1010 method without limiting to the evaluation method. For
example, the configuration changes needs human work
man-hour, and has a means for measurement and
prediction therefor, the work man-hour may be included
in metrics Ml.
1015 Column T904 includes a level of individual
metrics M2 for evaluating performance change with
configuration changes. Performance is an important
metrics in determining configuration of log database
3 . The reason is that the performance influences time
1020 required for monitoring work, and therefore a
predetermined or higher level of performance is needed
in accomplishing the monitoring work. For example, a
restoration time target during fault occurrence is
generally determined in service monitoring. During
1025 the fault occurrence, log database 3 is utilized for a
42
cause analysis. Therefore, time required for fault
restoration depends on response time required for log
data search. Accordingly, a predetermined level of
response performance is required for log database 3 in
1030 order avoid hindering the monitoring work. In an
example in the present embodiment, a response time
increase and decrease amount before and after the
configuration changes is used as one example of
metrics M2. An average of response time in a server
1035 per N data of search hit data counts is used for the
response time. Hereinafter, the response time per N
data of the search hit data counts is to be referred
to as criterial response time for convenience. In
column T904, conditions to be met by a metrics value
1040 applied to evaluating metrics M2 are configured. For
example, row T950 presents that a level of metrics M2
is configured to "1 second or more in a decrease of
criterial response time with the configuration
changes" in a certain individual policy. In the
1045 present case, configuration at which the response time
is not improved by 1 second or more is judged to be
configuration that does not meet the policy.
The description above is taken as one example,
and metrics M2 is evaluated by a method without
1050 limiting to the evaluation method. For example.
43
1055
1060
response time including processing in a client, and a
network, as well as the response time in the server,
may be used as the response time used for the metrics.
Moreover, a CPU utilization rate or the like but not
the response time may be used as the increase and
decrease amount of performance. Alternatively, if a
means for measurement and prediction exists, execution
time of monitoring work by the operations manager may
be applied as the metrics.
Column T905 includes a level of individual
metrics M3 for evaluating data volume change with
configuration changes. The data volume is an
important metrics in determining configuration of log
database 3. The reason is that cost is further
1065 required for maintenance and management as a log data
volume increases. In an example in the present
embodiment, a conservation data increase and decrease
volume before and after the configuration changes is
used as one example of metrics M3. A data
1070 conservation amount predicted when log data for a
configured retention is accumulated is used for the
data conservation amount. Hereinafter, the data
conservation amount is to be referred to as a
criterial data conservation amount for convenience.
1075 Conditions to be met by the metrics value applied to
44
evaluating metrics M3 are configured in column T904.
For example, row T951 presents that a level of metrics
M3 is configured to "1 GB or more in reduction of
criterial data conservation amount with the
1080 configuration changes" in a certain individual policy.
In the present case, configuration at which the data
conservation amount is not improved by 1 GB or more is
judged to be configuration that does not meet the
policy.
1085 The description above is taken as one example,
and metrics M3 is evaluated by a method without
limiting to the evaluation method. For example, a
data conservation amount per unit of time may be used.
In addition, in the present embodiment, only a
1090 configuration of log database 3 that meets all of the
three levels of metrics Ml, M2 and M3 of individual
policies is regarded as a configuration meeting the
individual policy. For example, the configuration of
log database 3 meeting the policy in row T951 is a
1095 configuration meeting "less than 60 minutes in time
required for the configuration changes," "less than 1
second in an increase of criterial response time with
the configuration changes" and "1GB or more in
reduction of criterial data conservation amount with
1100 the configuration changes."
45
1105
1110
Figure 11 is a diagram showing an example of
entire policy information T800 of configuration
changes. Entire policy information T800 includes data
describing a level that each tenant requires for the
entire configuration changes upon the configuration
changes of log database 3. In the present embodiment,
when configuration changes proposal generating unit 82
creates the configuration changes proposal, the unit
82 judges whether the configuration changes proposal
as a whole meets the level, and when the level is met,
presents the proposal. In the present embodiment,
only one of entire policy information T800 is to be
configuration for every tenant. Column T801 presents
ID for uniquely specifying entire policy information
T800. Column T802 includes a name of the tenant who
configures the policy. Columns T802 to T805 includes
a metrics level that a tenant requires for the
configuration changes as the whole. In the present
invention, three kinds of metrics values for
1120 evaluating the changes proposal as the whole are
calculated upon generating the configuration changes
proposal. When the calculated metrics values meet
levels presented in columns T803 to T805, the
configuration is adopted. Hereinafter, the metrics
1125 evaluating the changes proposal as the whole is to be
1115
46
1130
referred to as an entire metrics.
Column T803 includes a level of entire metrics M4
for evaluating time required for configuration
changes. The metrics corresponds to metrics Ml of
individual policy information, and the content is
almost same therewith, but different therefrom in
evaluating the time by gross time required for all
configuration changes.
Column T804 includes a level of entire metrics M5
1135 for evaluation performance change with configuration
changes. The metrics corresponds to metrics M2 of
individual policy information, and the content is
almost same therewith, but different therefrom in
evaluating the performance change by all of the
1140 performance of the configuration changes. More
specifically, in an example in the present embodiment,
an average of criterial response time regarding all
configuration after change is used as metrics M5.
Column T805 includes a level of entire metrics M6
1145 for evaluating data volume change with configuration
changes. The metrics corresponds to metrics M3 of
individual policy information, and the content is
almost same therewith, but different therefrom in
1150
evaluating the change by a total value of data
conservation amount of all log data after change
47
^R
1155
1160
In addition, in the present embodiment, the
configuration changes proposal need to meet all of the
three levels of metrics M4, M5 and M6 of the entire
policy in a manner similar to the individual policy.
For example, the configuration changes proposal of
tenant A presented in row T850 need to meet "less than
300 minutes in a total of time required for the all
configuration changes" and "less than 10 seconds in an
average of criterial response time regarding the all
configuration after the change," and "less than 100 GB
in a data conservation amount of the all log data
after the change." Configuration changes proposal of
tenant B presented in T851 need to meet "less than 300
minutes in a total of time required for all
configuration changes," and "less than 5 seconds in an
average of criterial response time regarding the all
configuration after the change" and "less than 50GB in
a data conservation amount of all log data after the
change."
The entire metrics used in the description above
is taken as one example and application of the present
invention is not limited to the evaluation methods.
In the present embodiment, entire policy
information T800 is to be applied by one for one
1175 tenant, but a plurality of entire policy information
1165
1170
48
^B
1180
1185
T800 may exist for one tenant in a manner similar to
individual policy information T900.
In the present invention, individual policy
information T900 and entire policy information T800
are configured in a unit of the tenant, but the
embodiment according to the present invention is not
limited to the present method. For example, the
policy information may be categorized and configured
not only for every tenant but also for every data type
name or instance name. Alternatively, the policy
information may be configured for every data
conservation format. For example, a policy for every
data type is determined in individual policy
information T900, and then a total value of a
1190 plurality of instance metrics values may be used for a
level value (example: less than 10 GB in a total value
of criterial data conservation amount of all the
instances of a certain data type).
Moreover, the policy information may be
configured by explicit input from the operations
manager, or may be automatically configured
mechanically.
The present invention also allows generation of
the configuration changes proposal without using
policy information at all, as long as the metrics
1195
1200
49
^B value is calculated. For example, the configuration
changes proposal may be generated by preliminarily
determining what metrics has priority, and
sequentially selecting better configuration having a
1205 high priority metrics value.
Figure 12 shows a sequence of configuration
changes proposal generation processing in a first
embodiment according to the present invention.
Configuration management interface 81 of log
1210 configuration management server 8 browses, upon
request of configuration review request from
operations manager terminal 7 of a tenant,
configuration information of log database 3 of the
tenant managed by the operations manager to perform
1215 configuration changes. Configuration changes proposal
generating unit 82 of log configuration management
server 8 (data configuration information generating
unit 83, data configuration information adjusting unit
84, index configuration information generating unit
1220 85, index configuration information adjusting unit 86
and configuration changes proposal displaying unit 87)
is invoked on occasion of the request from the
operations manager through configuration management
interface 81, uses the information stored in
1225 management database 4 to analyze a utilization
50
1235
situation of log database 3 of the tenant to
automatically generate the configuration changes
proposal in conformity with the situation and to
display the changes proposal on a screen of operations
1230 manager terminal 7 of the tenant. The changes
proposal approved by the tenant manager is stored in
configuration changes proposal database 5.
Subsequently, a detailed flowchart for an
individual unit in Figure 12 is described. In
addition, information to be generated in connection
with configuration changes proposal generation
processing and to be stored in configuration changes
proposal database 5 is also described in detail.
Figure 13 shows a flowchart of data configuration
information generation processing by data
configuration generating unit 83. In the processing,
raw data configuration information and compressed data
configuration information are generated.
Data configuration generating unit 83 acquires,
upon configuration review request from an operations
manager through configuration management interface 81,
operation history T500 of log monitoring unit 63
during designated tenant review target period (SlOl).
Here, in the present embodiment, an example is
1250 described in which the review target period is assumed
1240
1245
51
to be a period from time on which the configuration
changes is executed in previous time to current. The
period is assumed to be about one month. However, the
embodiment according to the present invention is not
1255 limited to the review target period. For example, if
each history is conserved for every configuration
changes, review for a long period of time over a
plurality of configuration changes can also be
executed in a manner similar to the operations in the
1260 first embodiment.
Next, data configuration generating unit 83
enumerates a unique combination of monitoring targets
(tenant, data type name and instance name) from
operation history T500 acquired in SlOl (S102) . On
1265 operation history T500, the tenant can acquire the
data type name from column T502, and the instance name
from column T506.
Subsequently, data configuration generating unit
83 repeats processing in S103 to S114 for the
1270 monitoring targets enumerated in S102.
Data configuration generating unit 83 selects
untested monitoring target S_KEY from the monitoring
targets enumerated in S102, and acquires group GRP of
a history having information presenting the monitoring
1275 target from operation history T500 (S103) . Then, data
52
^ ^ configuration generating unit 83 divides the GRP into
two groups: compressible group GRP_COM and noncompressible
group GRP_RAW (S104). As a judging
method thereof, for example, if access T510 to
1280 compressed data is "present", and response time T507
required for search is less than threshold Tl in
operation history T500, the operation history T500
(search target thereof) is judged to be compressible
data. Threshold Tl may be preliminarily determined
1285 individually (example: 20 seconds), or may be
otherwise. For example, calculation may be made by
another means or a level of metrics M5 of entire
policy information T800 may be used as Tl. Provision
of a threshold herein is made for the purpose of
1290 avoiding a situation of excessively poor response time
by searching the compressed data. For the purpose,
discrimination may be made only by existence or notexistence
of access to the compressed data without
limiting the time by the threshold. Moreover, a
1295 method may also be applied in which the all operation
histories are used for generation of raw data
configuration information without performing the
judgment per se described above.
In S105 to S109, raw data configuration
1300 information is generated based on operation history
53
•
T500 of GRP_RAW. Configuration in conformity with
actually accessed data can be generated by performing
the processing. First, data configuration generating
unit 83 generates configuration information CONF_RAW
1305 of raw data of monitoring target S_KEY (S105) . Here,
as an initial value of CONF_RAW, the unit 83
designates an expression (monitoring target = S_KEY,
data conservation format = raw data, and retention =
[unlimited, to current]). The designation presents a
1310 configuration state in which the retention is not
particularly limited. Next, data configuration
generating unit 83 calculates log utilization day
count TR_RAW for every history included in GRP_RAW
(S106) . Here, TR_RAW is determined by an expression
1315 (minimum value of log measurement time designated on
search conditions - access time) of operation history
T500. The calculation allows determination of access
to log data in any old period starting from the access
time. Then, data configuration generating unit 83
1320 calculates maximum day count TRM_RAW among TR_RAW of
each history calculated in S106 (S107). Then, data
configuration generating unit 83 configures an
expression [from TRiy[_RAW days before to current] to
retention of CONF_RAW (S108) . Finally, data
1325 configuration generating unit 83 registers CONF_RAW as
54
generated configuration information (S109) . In
addition, a case where no history exists in GRP_RAW
can be considered. In the present case, processing in
S105 to S109 is omitted.
1330 In SllO to S114, compressed data configuration
information is generated based on operation history
T500 of GRP_COM. A general processing flow is similar
to the flow of raw data configuration information
generation in S105 to S109. Configuration in
1335 conformity with actually accessed compressible data
can be generated by performing the processing. First,
data configuration generating unit 83 generates
compressed data configuration information CONF_COM of
monitoring target S_KEY (SllO) . Here, as an initial
1340 value of CONF_COM, the unit 83 designates an
expression (monitoring target = S_KEY, data
conservation format = compressed data, retention =
[unlimited, to current]). The designation presents a
configuration state in which the retention is not
1345 particularly limited. Next, data configuration
generating unit 83 calculates log utilization day
count TR_COM for every history included in GRP_COM,
and calculates the maximum day count TRM_COM (Sill to
S112). The calculation method is similar to the
1350 method for TR RAW and TRM COM. Then, data
55
configuration generating unit 83 configures an
expression [from TRM_COM days before to TRM_RAW days
before] to retention of CONF_COM (S113) . Here,
configuration of an end time point of the retention to
1355 a start time point of the raw data configuration
retention is made for the purpose of continuing the
data retention. Finally, data configuration
generating unit 83 registers generated CONF_COM as
configuration information (S114) . In addition, a case
1360 where no history exists in GRP_COM can be considered
in the processing. In the present case, processing in
SllO to S114 is omitted.
Figure 14 is a diagram showing an example of
configuration information T1200 after being changed
1365 upon configuration review request from an operations
manager. Column T1201 includes ID for uniquely
identifying a log database configuration. Columns
T1202 to T1204 include information presenting a
monitoring target. Column T1205 presents a data
1370 conservation format to be designated for the
monitoring target. Column T1206 presents a target
data column to be indexed. Column T1207 includes data
retention to be configured.
Configuration information T1200 after the change
1375 is generated by configuration changes proposal
56
generating unit 82, and includes the configuration
meeting the entire metrics and the individual metrics
of the policy. Row T1250 includes information
presenting the monitoring target being raw data, and
1380 row T1251 includes a compressed data retention
configuration of the monitoring target. Rows T1252 to
T1254 include an index configuration to raw data
having information presenting the monitoring target
identical with the target in row T1250. The content
1385 of the column constituting the information is similar
to the content of current configuration information
shown in Figure 8, and therefore the description is
omitted.
In compressed data configuration information
1390 generation processing in the present embodiment, the
start time point of the retention is calculated based
on the access situation, but the start time point may
be configured to "unlimited" without performing the
calculation. A risk of excessive data deletion can be
1395 suppressed by configuring no limit. In the present
case, compressed data are not automatically deleted,
and are deleted only by operation by the operations
manager.
Moreover, in data configuration information
1400 generation processing in the present embodiment, the
57
configuration information retention is designated to
the maximum value of the expression (the minimum value
of log measurement time designated on the search
conditions - access time) of operation history T500,
1405 but a value obtained by multiplying a safety factor
(example: 1.2 times) to the maximum value may be used.
The process is effective in the case where use of the
log is just started, and the utilization day count may
increase, or the like.
1410 Figure 15 shows a flowchart of data configuration
information adjustment processing by data
configuration adjusting unit 84. Data configuration
adjusting unit 84 judges, after data configuration
information generation processing shown in Figure 13
1415 finishes, whether generated configuration information
T1200 meets a configuration changes policy to adjust
generated configuration information T1200, thereby
generating final configuration changes proposal. The
flow is described below.
1420 First, data configuration adjusting unit 84
acquires current configuration information T600
(Figure 8) from management database 4, and generated
configuration information T1200 (Figure 14) from data
configuration generating unit 83, respectively (S201).
1425 Next, data configuration adjusting unit 84 enumerates
58
1435
a unique combination of monitoring targets (tenant,
data type name and instance name) from current
configuration information T600 and generated
configuration information T1200 as acquired in S201
1430 (S202) .
Subsequently, data configuration adjusting unit
84 repeats processing in S203 to S215 for the
monitoring targets enumerated in S202.
Data configuration adjusting unit 84 selects
untested monitoring target S_KEY from the monitoring
targets enumerated in S202 (S203). Then, data
configuration adjusting unit 84 acquires raw data
configuration information A and compressed data
configuration information A in agreement with S_KEY
from current configuration information T600 (S204) .
Moreover, the unit 84 acquires raw data configuration
information B and compressed data configuration
information B in agreement with S_KEY from generated
configuration information T1200 in a similar manner
(S205) .
1440
1445
1450
Next, when retention of generated raw data
configuration information B is shorter than the
retention of current raw data configuration
information A, data configuration adjusting unit 84
generates another candidate C of compressed data
59
^R configuration information so as to retain compressed
data for a period of the difference (S206 to S207) .
The process is applied for relieving generation of
excessive data deletion based on reducing the raw data
1455 retention. The unit 84 provides a configuration value
of compressed data configuration information C with an
expression (monitoring target = S_KEY, data
conservation format = compressed data, retention =
[from time on which retention starts in current raw
1460 data configuration information B to time on which
retention starts in generated raw data information
A] ) .
Furthermore, data configuration adjusting unit 84
finds out a best configuration combination from the
1465 raw data configuration information and the compressed
data configuration information as presented above, and
registers the combination as configuration information
of monitoring target S_KEY after change (S208 to
S215) . First, the unit 84 generates six combinations
1470 (hereinafter, configuration candidate pair) of raw
data configuration information and compressed data
configuration information as presented in S204 to S207
as described below (S208) .
An expression (configuration candidate pair = (A,
1475 a) (A, b) (A, c) (B, a) (B, b) (B, c)).
60
Then, data configuration adjusting unit 84
repeats processing in S209 to S212 for the
configuration candidate pair in S208.
Data configuration adjusting unit 84 selects an
1480 untested pair (raw data configuration information X,
compressed data configuration information x) from the
configuration candidate pair in S208 (S209) . Then,
data configuration adjusting unit 84 configures a
start time point of the retention of raw data
1485 configuration information X at a end time point of the
retention of compressed data configuration information
X (S210). The configuration includes processing for
allowing continuation of the retention in the
compressed data configuration and the raw data
configuration. Next, data configuration adjusting
unit 84 judges a change category of raw data
configuration information X and compressed data
configuration information x (S211) . The unit 84
judges the category as a change category "addition"
when current configuration information is empty, a
change category "deletion" when configuration
information (X or x) selected from the configuration
candidate pair is empty, and a change category
"change" when both are not empty. Furthermore, in the
1500 case of the change category "change," when the
1490
1495
61
1505
configuration information is completely in agreement
therewith, no change is made, and therefore subsequent
processing is omitted.
Next, data configuration adjusting unit 84
calculates values of metrics Ml, M2, and M3 of raw
data configuration information X and compressed data
configuration information x, respectively (S212).
Subsequently, one example of the calculation method of
each metrics value is given.
1510 Assumed processing time required for the
configuration changes is used for metrics Ml in the
present embodiment as described above. An example of
the calculation method for the processing time is
presented. In the present embodiment, a performance
1515 unit price and a trend for every category or every
target of the configuration changes are to supposedly
be known. The performance unit price, current
configuration information T600, configuration
information after the change (configuration
1520 information selected from the configuration candidate
pair, herein), and data volume information T700 are
used, and calculation is to be made according to
expressions 1-1 and 1-2 described below.
Assumed processing time required for
1525 configuration changes =
62
^B
1535
performance unit price required for configuration
changes per one data count x data count to be
processed (expression 1-1).
Data count to be processed =
1530 conservation data count T707 of configuration ID
T701 identical with the configuration information in
data volume information T700
X ABS (a day count of retention of configuration
information after change)
- a day count of retention of current
configuration information) -^ a day count of the
retention of the current configuration information
(function ABS() refers to a function for calculating
an absolute value in a parenthesis, herein)
(expression 1-2).
An example is presented in which actual values
are applied to the calculation expressions. First, a
performance unit price of deletion processing per
count is assumed to be 1 millisecond/data in raw data
1545 deletion processing. Moreover, the day count of the
retention of the current configuration information is
assumed to be 100 days, and the day count of the
retention of the configuration information after the
change is assumed to be 80 days. Furthermore, the
1550 conservation data count is assumed to be 10 million
1540
63
^^m
counts. In the present case, assumed processing time
required for the configuration changes is calculated
to be "[1 millisecond/data x 10 million counts x ABS
(80 days - 100 days) -:- 100 days = 2,000 seconds = 33
1555 minutes."
In the present embodiment, the example is given
in which the performance unit price is known, but the
present invention is restricted thereto. For example,
when a performance unit price is unknown, addition or
1560 deletion processing is performed on actual equipment
using a small volume of log data for mechanically
understanding a performance unit price, and then the
performance unit price may be determined from the
measuring results for the time. Moreover, when the
1565 trend is unknown, a trend of a past configuration
changes track record may be analyzed by applying a
multivariate analysis technique. For example, when
processing time depends on a data volume during
retention, a prediction calculation expression is
1570 determined by conducting a statistical analysis of the
past configuration changes track record, and the
expression may be applied to calculation of metrics
Ml .
In the present embodiment, as described above,
1575 the change amount of the reference response time in
64
^1^ connection with the configuration changes is used for
metrics M2. In the present embodiment, an analysis of
response time after the change is assumed to be
allowed from a trend of a past operation history, and
1580 is calculated using operation history T500, current
configuration information T600, configuration
information after the change (configuration
information selected from the configuration candidate
pair herein), and data volume information T700
1585 according to expressions below.
Change amount of criterial response time =
criterial response time of configuration
information after change - criterial response time of
current configuration information (expression 2-1).
1590 Criterial response time of current configuration
information =
an average of criterial response time of the all
operation history of a monitoring target identical
with the target of the configuration information
1595 (expression 2-2).
Criterial response time of an operation history =
response time required for search x (N data in a
data count serving as a criterial -^ data hit in
search) (expression 2-3).
1600 Criterial response time of configuration
65
•
information after change =
Criterial response time of current configuration
information x performance improvement coefficient by
change (expression 2-4).
1605 Performance improvement coefficient by the change
func (data row counts of indices to be used after
change, access to compressed data after the change)
(expression 2-5).
1610 Here, the performance improvement coefficient by
the change is to be expressed using a numeric value of
0 to 1, and as the numeric value is smaller, the
performance is presented to be better. Function func
() for determining the coefficient includes a
1615 prediction expression determined by a statistical
analysis (for example, multiple regression analysis)
of a performance trend using as input parameters the
data row counts of indices used for search of the past
operation history, and information on existence or
1620 not - existence of access to compressed data.
The calculation method for metrics M2 in the
present embodiment is taken as one example, and
therefore any other calculation method may be applied.
For example, another machine learning technique such
1625 as a neutral network may also be applied.
66
•
As described above, the change volume of data
conservation amount before and after the configuration
changes is used for metrics M3 in the embodiment. In
the embodiment, a analysis of the data conservation
1630 amount after the change is to be allowed from a trend
of data volume information T700, and is calculated
using current configuration information T600,
configuration information after the change
(configuration information selected from the
1635 configuration candidate pair herein), and data volume
information T700 according to expressions as described
below.
Change volume of criterial data conservation
amount =
1640 data conservation amount T706 having
configuration ID T701 identical with the configuration
ID of configuration information in data volume
information T700
X ABS (day count of retention of configuration
1645 information after change - day count of retention of
current configuration information
-;- a day count of retention of current
configuration information (expression 3-1).
In the calculation described above, when a
1650 category of configuration changes is addition, data
67
1655
1660
conservation amount T706 having configuration ID T701
identical with the configuration ID of configuration
information in data volume information T700 does not
exist. In the present case, the change amount is to
be presumed from an alternative means described below.
In order to allow presumption herein, in the present
embodiment, a compression ratio for converting raw
data into compressed data, and a ratio of a data
volume in comparison with the raw data in the case of
generating an index are assumed to be known. When the
data conservation amount of the identical monitoring
target exists, a data volume is acquired using the
numeric value, and a value obtained by multiplying the
numeric value with the compression ratio and the ratio
of data volume as described above is used. Moreover,
when the data conservation amount of the identical
monitoring target does not exist at all, an average of
the data conservation amount of other all other
monitoring targets is to be used.
1670 The calculation method of metrics M3 in the
present embodiment is taken as one example, and
therefore any other calculation method may be applied.
For example, if the retention is also calculated upon
calculating data volume information T700 and the
1675 calculated period is taken into consideration during
1665
68
•
calculating metrics M3, accuracy of evaluation can be
further improved.
After calculation of the individual metrics
finishes by the method described above, data
1680 configuration adjusting unit 84 acquires individual
policy information T900 corresponding to monitoring
target S_KEY (S214) . Then, among the all
configuration candidate pairs, configuration
information meeting the individual policy conditions,
1685 and having a best metrics value is registered as
configuration information T1200 after the change.
Moreover, the unit 84 registers current configuration
information T600, configuration information T1200
after the change and values of metrics Ml, M2 and M3
1690 values as difference information of configuration
changes (S215) . Here, how to select configuration
information having the best metrics value includes,
for example, a method for sorting configuration
candidate according to priority of a data conservation
1695 format and an individual metrics as preliminarily
determined (example: raw data configuration
information takes priority as the data conservation
format, and also the metrics takes priority in the
order of M2, M3 and Ml) and selecting the most
1700 significant configuration information.
69
^B As a result of repeating the operations in S203
to S216 as described above, configuration information
T1200 after the change and difference information
TllOO of the configuration changes are registered.
1705 Then, data configuration adjusting unit 84
distinguishes whether the configuration changes meets
the entire policy. First, the unit 84 acquires entire
policy information T800 corresponding to monitoring
target S_KEY (S217) . Then, the unit 84 calculates
1710 entire values of metrics M4, M5 and M6 values from
difference information TllOO of the configuration
changes (S218) . Hereinafter, one example of the
calculation method of each metrics value is given.
Metrics M4 in the present embodiment is evaluated
1715 by gross time required for the all configuration
changes. A calculation method is almost similar to
the method presented in metrics Ml in S212. However,
in metrics M4, gross time required for the all
configuration changes is calculated. Therefore, a
1720 total value of assumed processing time (determined
according to expression 1-1) required for the all
configuration changes is calculated, and a result
thereof is set as a value of metrics M4.
A calculation method of metrics M5 in the present
1725 embodiment is almost similar to the method presented
70
in metrics M2 in S212. However, in metrics M5, an
average of criterial response time for the all
configuration after the change is calculated.
Therefore, first, criterial response time of
1730 configuration information after the change is
calculated for configuration information T1200 after
the all change using expressions 2-4. Then, an
average of determined all criterial response time is
calculated, and a result thereof is set as a value of
1735 metrics M5. Here, "index data column to be used after
change" and "existence or not - existence of access to
compressed data after change" serving as input values
of expressions 2-5 are to be estimated by comparison
between search conditions T504 to T506 in operation
1740 history T500 and configuration information T1200 after
the change. For example, configuration information
T1200 after change in which the data columns used in
search conditions T504 to T506 in operation history
T500 and a data column of index target data T600 in
1745 configuration information T1200 after the change are
most agreed is regarded as the index data column to be
used in search of operation history T500 thereof.
A calculation method of metrics M6 in the present
embodiment is almost similar to the method presented
1750 in metrics M3 in S212. However, in metrics M6, a
71
1755
1760
total value of a data conservation amount of all log
data after change is calculated. Therefore, an
individual data conservation amount of configuration
information T1200 after the change is calculated using
the following expression 3-2 obtained by correcting
expression 3-1, and a total value is determined.
Criterial data conservation amount =
data conservation amount T706 having
configuration ID T701 identical with the ID of
configuration information in data volume information
T700
X a day count of retention of configuration
information after change -H a day count of retention of
current configuration information (expression 3-2).
Next, data configuration adjusting unit 84
compares between entire policy information T800 and
calculated entire values of metrics M4, M5 and M6, and
distinguishes whether configuration information T1200
after the change meets a level (S219). When the level
1770 is met, the unit 84 registers M4, M5 and M6 as entire
metrics value T1300 of the configuration changes
(S220) . When the level is not met, the unit 84
regards the changes proposal to be insufficient, and
deletes difference information T1200 of the
1775 configuration changes. As configuration information
1765
72
||P T1200 after the change, the unit 84 uses configuration
information T600 before the change as is (S221) .
In S221 in the present embodiment, when the
entire metrics value does not meet entire policy
1780 information T800, the unit 84 regards the metrics
value to be insufficient as the changes proposal to
terminate processing, and uses configuration
information T600 before the change as is. The reason
of terminating processing is for minimizing a
1785 calculation amount. Therefore, when consideration of
the calculation amount is not needed, a calculation
result up to S215 is left, another configuration that
is not selected in S215 is reregistered, and
recalculation may be made. Alternatively, an entire
1790 metrics value is calculated for combinations of all of
configurations that meet individual policy information
T900, and configuration having a best entire metrics
value may be adopted from the entire metrics values.
Moreover, judgment processing whether or not the
1795 entire policy is met is not executed individually, and
may be executed after processing finishes up to
adjustment processing of the below-mentioned index
configuration information.
Furthermore, all of metrics Ml to M6 are not
1800 needed to be used for judgment. For example, metrics
73
^ ^ M2 and M5 also depends on a later-determined index
configuration, and therefore are omitted in the data
configuration information adjustment processing, and
calculation and judgment of metrics M2 and M5 may be
1805 collectively executed upon index configuration
information adjustment processing.
Figure 16 is a drawing showing an example of
difference information TllOO of the configuration
changes. Configuration changes difference information
1810 TllOO is generated by configuration changes proposal
generating unit 82. In the present embodiment, only
configuration to be changed by the latest
configuration changes proposal is registered into the
difference information TllOO of the configuration
1815 changes. In columns TllOl to Tllll, configuration
information T600 before change (namely, current) and
configuration information T1200 after the change are
presented. Column TllOl includes ID for uniquely
identifying configuration of log database 3. Columns
1820 T1102 to T1104 include information presenting a
monitoring target. Column T1102 includes a name of a
tenant managing log data to be configured herein.
Column T1103 includes a name of a data type of log
data to be configured herein. Column T1104 includes a
1825 name of an instance outputting log data to be
74
unconfigured herein. Column T1105 presents a data
conservation format designated for the monitoring
target. Columns TllOl to T1105 correspond to columns
T601 to T605 of current configuration information
1830 T600, and columns T1201 to T1205 of configuration
information T1200 after the change. Columns T1107 and
T1108 include index target data and retention in
configuration information before the change,
respectively, and correspond to columns T606 to T607
1835 of current configuration information T600. When
configuration is categorized into addition,
corresponding current configuration information T600
does not exist, and therefore a symbol "-" is
registered into the columns. Columns TlllO and Tllll
1840 include index target data and retention in
configuration information T1200 after the change,
respectively, and correspond to columns T1106 to T1107
of current configuration information TllOO. If
configuration is categorized into deletion,
1845 corresponding configuration information T1200 after
the change does not exist, and therefore a symbol "-"
is registered into the columns. Column T1112 includes
a change category of the configuration changes. Any
one of values of "change," "deletion" and "addition"
1850 is registered into the change category. Columns T1113
75
^^ to T1115 include metrics values that are applied to
evaluating individual metrics Ml to M3 of the
configuration. A specific example is described using
row T1150. Row T1150 presents changing a start time
1855 point of retention from "180 days before" to "200 days
before" for raw data configuration of a monitoring
target (tenant A, Web event log of service 1,
aplserverl). Results of evaluating the configuration
changes present 10 minutes required for the
1860 configuration changes (column T1113), a 1 second
decrease in criterial response time (column T1114),
and a 1 GB decrease in criterial data conservation
amount (column T1115) . Configuration changes
difference information is registered also for rows
1865 T1151 to T1154 in a similar manner.
Figure 17 shows a flowchart of index
configuration information generation processing by
index configuration information generating unit 85 in
the first embodiment. A general flow and the
1870 processing content partially executed in the flow are
similar to the general flow and the content of the
flowchart of data configuration information generating
unit 83. Therefore, the description is omitted for a
part of similar content.
1875 Index configuration information generating unit
76
85 acquires, upon configuration review request from an
operations manager through configuration management
interface 81, operation history T500 of log monitoring
unit 63 during a designated tenant review target
1880 period (S301) . The processing is similar to the
processing in SlOl of data configuration generating
unit 83.
Next, index configuration information generating
unit 85 enumerates a unique combination of monitoring
1885 target + index target data (tenant, log data type
name, instance name, index target data) from operation
history T500 acquired in S301 (S302) . As the index
target data, the unit 85 uses index T509 used in
search on operation history T500. Moreover, in the
1890 case of operation history T500 in which an index is
not used for search, the unit 85 designates, as an
exception, "no target data" into index target data.
Subsequently, index configuration information
generating unit 85 repeats processing in S303 to S319
1895 for the monitoring target + index target data as
enumerated in S302.
Index configuration information generating unit
85 selects untested data I_KEY from the monitoring
target + index target data as enumerated in S302, and
1900 acquires group I_GRP of a history having the
77
monitoring target + index target data acquired from
operation history T500 (S303).
Then, index configuration information generating
unit 85 generates a candidate of new index
1905 configuration information corresponding to I_KEY (S304
to S316) . First, index configuration information
generating unit 85 generates list L for temporarily
retaining a generated index configuration information
candidate (S304) . Next, index configuration
1910 information generating unit 85 acquires individual
policy information T900 corresponding to I_KEY (S305) .
Furthermore, the unit 85 generates power set of the
data column designated on search conditions (columns
T504 to T506) from operation history T500 included in
1915 I_GRP (S306) . A method in S305 is described by taking
row T553 of operation history T500 in Figure 7 as an
example. In the row, "log measurement time" is
designated in columns T504 to T505, and "data type
name" and "instance name" in column T506, and thus
1920 three data columns in total are designated for search.
Thus, as the power set of the data column, {{no
target data}, {log measurement time}, {data type
name}, {instance name}, {log measurement time, data
type name}, {log measurement time, instance name},
1925 {data type name, instance name}, {log measurement
78
time, data type name, instance name}} is generated.
Then, index configuration information generating
unit 85 repeats processing in S308 to S316 for an
element of the power set of the data column generated
1930 in S305.
Index configuration information generating unit
85 selects untested element PARAM_SET of the power set
of the data column (S308) . Further, index
configuration information generating unit 85 generates
1935 candidate CONF_IDX of index configuration information
of I_KEY using the PARAM_SET (S309) . In the initial
configuration value, the unit 85 designates an
expression (monitoring target = monitoring target of
I_KEY, conservation format = index, index target data
1940 = PARAM_SET, retention = [unlimited, to current].
Furthermore, index configuration information
generating unit 85 calculates a log utilization day
count of each operation history T500 included in I_GRP
(minimum value of log measurement time designated on
1945 search conditions - access time), and sorts the day
count in ascending order (S310) . Then, index
configuration information generating unit 85 judges
whether untested data are in the log utilization day
count calculated in S310 (S311), and if no, the unit
1950 85 regards index configuration information using the
79
PARAM_SET to be insufficient, and returns to
processing for selecting next PARAM_SET (S308) . If
any, the unit 85 selects untested and narrowest log
utilization day count TR_IDX from the log utilization
1955 day count in S311 (S312) . Furthermore, the unit 85
configures the log utilization day count TR_IDX as
retention of candidate CONF_IDX of index configuration
information (S313) . Furthermore, the unit 85
calculates values of metrics Ml, M2 and M3 of
1960 candidate CONF_IDX of index configuration information
(S314) . The calculation method applies a method
similar to the above-mentioned method. The unit 85
judges whether calculated metrics Ml, M2 and M3 meet
the individual policy and if met, registers candidate
1965 CONF_IDX of the index configuration information into
list L (S316). If not met, the unit 85 reselects new
TR_IDX (S312) .
The unit 85 lists candidates of index
configuration information of I_KEY by processing in
1970 S304 to S316 as described above, and then selects one
of the candidates and registers the candidate as
generated configuration information. For the purpose,
index configuration information generating unit 85
first extracts configuration information expected to
1975 be least in a data volume from list L (S317) .
80
1980
1985
1990
1995
Specific examples of the method include a method for
selecting configuration information in which a product
of an index target data count multiplied by retention
is minimized. Processing in S317 includes one example
for selecting the best candidate of the list, and the
best candidate may be elected by any other method or
criteria. Then, index configuration information
generating unit 85 registers the configuration
information extracted in S317 as generated
configuration information (S318) .
Index configuration information generating unit
85 repeats processing in S303 to S319 described above,
and generates index configuration information, and
then eliminates unnecessary index configuration
information. For the purpose, the unit 85 compares
between the index target data column and the retention
of the generated configuration information with each
other (records with each other), and confirms an
inclusion relation (S320) . An index to be included
can be alternated by an included index, and therefore
is unnecessary. Consequently, the unit 85 deletes all
configuration information included in any other
configuration information from the generated
configuration information (S322) . Thus, the data
2000 volume can be further saved.
81
In the configuration information generation
processing in the present embodiment, metrics Ml to M6
are all used for judgment. However, all of the
metrics are not necessarily used. For example, when
2005 metrics M2 and M5 are important, calculation and
judgment of the metrics only may be performed.
Figure 18 shows a flowchart of index
configuration information adjustment processing by
index configuration information adjusting unit 86 in a
2010 first embodiment. A general flow of the processing
and the processing content partially executed in the
flow are similar to the flow and the content in the
flowchart of data configuration information adjusting
unit 87. Therefore, the description on a part similar
2015 in the content is omitted.
Index configuration information adjusting unit 86
judges, after the index configuration information
generation processing by index configuration
information generating unit 85 shown in Figure 17
2020 finishes, whether the generated configuration
information meets the configuration changes policy,
and adjusts generated configuration information,
thereby generating final configuration changes
proposal. A flow thereof is described below.
2025 First, index configuration information adjusting
82
2030
2035
unit 86 acquires current configuration information
T600 from management database 4, and configuration
information generated by processing by index
configuration information generating unit 85,
respectively (S401). Next, index configuration
information adjusting unit 86 ties current
configuration information T600 acquired in S401, and
generated configuration information, and creates a
pair before and after configuration changes (S402) .
One example of the method is given. First, the unit
86 applies as a configuration changes pair in
agreement with a monitoring target in a data
conservation format and an index target data.
Furthermore, when the data conservation format agrees
2040 with the format of the monitoring target and the index
target data does not agree therewith, the unit 86
applies, as the configuration changes pair,
configuration information in which the agreed data
column count is largest with each other. For an
2045 otherwise configuration, the unit 86 regards the
configuration as addition or deletion, and applies as
a configuration changes pair in which configuration
information before and after change is empty. Next,
index configuration information adjusting unit 86
2050 enumerates a unique combinations of monitoring target
83
^L + index target data (tenant, data type name, instance
name, index target data) from the pair before and
after the configuration changes in S402 (S403) .
Subsequently, index configuration information
2055 adjusting unit 86 repeats processing in S404 to S413
for the monitoring target + index target data
enumerated in S403.
Index configuration information adjusting unit 86
selects untested target I_KEY from the combination of
2060 the monitoring target + index target data in S403
(S404) . Next, index configuration information
adjusting unit 86 acquires index configuration
information CURRENT corresponding to I_KEY from
current configuration information T600, and index
2065 configuration information CREATED corresponding to
I_KEY from the generated configuration information,
respectively (S405) . Then, index configuration
information adjusting unit 86 judges a change category
of CURRENT and CREATED (S406) . The unit 86 judges the
2070 category to be change category "addition" when CURRENT
is empty, change category "deletion" when CREATED is
empty, and change category "change" when both are not
empty.
Then, index configuration information adjusting
2075 unit 86 calculates values of metrics Ml, M2 and M3 of
84
index configuration information, respectively (S407) .
A calculation method therefor is as mentioned above.
Further, index configuration information adjusting
unit 86 acquires individual policy information T900
2080 corresponding to I_KEY (S408), and judges whether
calculated values of metrics Ml, M2 and M3 meet the
individual policy (S409). As a result of judgment,
when the values do not meet the individual policy, the
unit 86 regards CREATED to be insufficient as change,
2085 and registers CURRENT as configuration information
T1200 after the change (S412) . As a result of
judgment, when the values meet the individual policy,
the unit 86 registers CREATED as configuration
information T1200 after the change (S410).
2090 Furthermore, the unit 86 registers, as difference
information TllOO of the configuration changes (I_KEY,
CURRENT, CREATED, a change category. Ml, M2 and M3)
(S411) .
Index configuration information adjusting unit 86
2095 performs processing described above, and registers
configuration information TllOO after the change and
difference information TllOO of the configuration
changes, and then performs judgment based on entire
policy information T800. Index configuration
2100 information adjusting unit 86 calculates entire
85
^h metrics values M4, M5 and M6 from difference
information TllOO of the configuration changes (S414).
A calculation method therefor is as mentioned above.
Next, the unit 86 acquires entire policy information
2105 T800 corresponding to I_KEY (S415). Furthermore,
index configuration information adjusting unit 86
compares between calculated entire metrics values M4,
M5 and M6, and entire policy information T800, and
distinguishes whether configuration information T1200
2110 after the change meets a level (S416). When the
values meet the level, the unit 86 registers M4, MB
and M6 as entire metrics value T1300 of the
configuration changes (S417) . When the values do not
meet the level, the unit 86 regards the changes
2115 proposal to be insufficient, and deletes difference
information T1200 of the configuration changes. As
configuration information T1200 after the change, the
unit 86 uses configuration information T600 before the
change as is (S418).
2120 In S418, when entire metrics value T1300 does not
meet entire policy information T800, the unit 86
regards the item to be insufficient as the changes
proposal to discontinue processing, and uses
configuration information T600 before the change as
2125 is. The reason of discontinuing processing is for
86
^^ decreasing a calculation amount. For the purpose,
when consideration of the calculation amount is not
needed, a calculation result of index generation
processing is left, and another configuration that is
2130 not selected in S317 is reregistered, and
recalculation may be made. Alternatively, entire
metrics value T1300 is calculated for combinations of
all of configurations that meet individual policy
information T900, and configuration having a best
2135 entire metrics value may be adopted from the entire
metrics values.
Figure 19 is a drawing showing an example of
entire metrics value T1300 of the configuration
changes. The diagram shows a result obtained by
2140 configuration changes proposal generating unit 82
evaluating an entire metrics of difference information
TllOO of the configuration changes. Column T1301
includes a name of a tenant who manages log data
subjected to configuration changes proposal. Columns
2145 T1302 to T1304 include metrics values of entire
metrics M4 to M6 of the configuration. For example,
entire metrics values of the configuration changes of
a tenant presented in row T1350 include "210 minutes
in a total of time required for the all configuration
2150 changes," "8 seconds in an average of criterial
87
^^ response time for all of configuration after change"
and "90 GB in a data conservation amount of all log
data after the change," and all meets the policy of
entire policy information T800 of the configuration
2155 changes .
In the adjustment processing of configuration
information in the present embodiment, metrics Ml to
M6 are all used for judgment. However, all of the
values may not be used for calculation.
2160 Moreover, processing in each flowchart described
above can also be executed without using individual
policy information T900 and entire policy information
T8000. For example, if another means for judging
superiority or inferiority of each metrics (priority
2165 of a metrics, or the like) exists, the method is
applied as an alternative, and the method according to
the present invention may also be executed.
Figure 20 is a drawing showing an example of a
configuration changes recommended screen display that
2170 configuration changes proposal display unit 87
presents to an operations manager. The screen
includes a screen for confirming the configuration
changes proposal requested by the operations manager,
and performing operation thereto. The recommended
2175 screen display 870 is constituted of entire metrics
88
^fc value display 871 of configuration changes difference
information display 872 of the configuration changes
proposal, approval or rejection selection button 873
of the configuration changes, configuration changes
2180 reflection button 874 and resubmitted proposal button
875. On entire index value display 871 of the entire
metrics value T1300 of the configuration changes as
previously evaluated is drawn. Moreover, on
difference information display 872 of the
2185 configuration changes difference information TllOO of
the configuration changes as previously generated in
which the metrics is evaluated is drawn. Moreover, an
increase and decrease of a charged amount from a
monitoring platform provider corresponding to each of
2190 difference information to a tenant manager is
simultaneously displayed. The operations manager can
confirm the configuration changes content and an
effect thereof, and an increase and decrease of cost
by browsing the displays. Approval or rejection
2195 selection button 873 of the configuration changes
includes interface for selecting "approval" or
"rejection" to be made for the change for every
configuration. The button is prepared for every row
of difference information TllOO (873-1 to 873-3).
2200 Then, symbols 874 and 875 refer to buttons for
89
^^ executing operation based on difference information
TllOO and selection information of approval or
rejection selection button 873. If configuration
changes reflection button 874 is pressed down, the
2205 button 874 updates configuration and current
configuration information T600 on log database 3 based
on difference information TllOO and the selection
information. If resubmitted proposal button 875 is
pressed down, the button 875 regenerates configuration
2210 changes proposal based on difference information TllOO
and selection information T1400. A method of the
resubmitted proposal is described in detail in a
second embodiment.
As described above, configuration changes
2215 proposal generating unit 82 according to the present
invention automatically generates the configuration
changes proposal using operation history T500, current
configuration information T600 and data volume
information T700. Moreover, the unit 82 presents
2220 metrics information to be a judgment source for the
effect and the influence by the configuration changes.
The operations manager can select configuration
changes based on the metrics information using
automatically generated configuration changes
2225 proposal. Therefore, the present invention is
90
^^ effective in allowing reduction of a man-hour of log
database configuration examination work. Accordingly,
the present invention is effective in allowing
reduction of cost required for monitoring by an
2230 operator.
Moreover, in the present invention, configuration
changes proposal are generated by analyzing a kind and
a period of log data actually utilized from operation
history T500. Thus, configuration in conformity with
2235 actually accessed data and period therefor can be
generated. The present invention is effective in
allowing ever-further reduction of unnecessary data on
the log database by performing the configuration
changes according to the proposal. Moreover, the
2240 present invention is effective in allowing suppression
of loss caused by excessive log data deletion (fault
correspondence man-hour increase, opportunity loss)
that has been so far generated. Accordingly, an
effect can also be expected in intensifying log
2245 monitoring of the monitoring platform by the
operations manager.
Furthermore, in the present invention,
configuration changes are generated using individual
policy information and entire policy information
2250 presenting a request level that the operations manager
91
^^ requires for the configuration changes, thereby
allowing recommendation of configuration changes
proposal further in conformity with a request from the
operational manager in comparison with a case where
2255 the information is not used. Therefore, the
configuration changes can be generated with higher
accuracy in comparison with the case where the
information is not used. Moreover, the request is
judged from two viewpoints of individual policy
2260 information T900 and entire policy information T800,
thereby allowing generation of configuration changes
proposal with higher accuracy in comparison with a
case where only one of the information is used. Thus,
recommendation with high accuracy is made, and thus an
2265 effect can be expected in allowing further reduction
of a man-hour of log database configuration
examination work by the operations manager.
In the present invention, tenant information T400
is used in generation of the configuration changes
2270 proposal. Thus, generation of the configuration
changes proposal using delicate operation history T500
or policy information for every tenant can be
performed. Therefore, accuracy of a configuration
changes proposal can be further improved in comparison
2275 with a case where the information is not used.
92
^k Therefore, the present invention is effective in
allowing further reduction of a man-hour of log
database configuration examination work by the
operations manager.
2280 Thus, cost required for log monitoring (in
particular, management of the log database) by the
operations manager of the monitoring platform can be
reduced by assisting the configuration changes work in
the present embodiment. In the present embodiment,
2285 first, the configuration changes proposal are
automatically generated. Metrics information to be a
judgment source for the effect and the influence by
the configuration changes is presented. Thus, a manhour
of log database configuration examination work
2290 can be reduced. In the present embodiment, second,
the configuration changes proposal are generated by
analyzing a kind and a period of log data actually
utilized from the operation history. Unnecessary data
on the log database can be reduced by performing the
2295 configuration changes according to the proposal.
Moreover, log monitoring of the monitoring
platform by the operations manager can be intensified.
In the present execution form, the configuration
changes proposal in conformity with utilization are
2300 generated. Therefore, loss caused by excessive log
93
^^ data deletion (fault correspondence man-hour increase,
opportunity loss) that has been so far generated can
be suppressed. Moreover, the configuration changes
proposal in conformity with utilization are generated.
2305 Therefore, loss caused by excessive log data deletion
(fault correspondence man-hour increase, opportunity
loss) that has been so far generated can be
suppressed.
In addition, in the first embodiment, an example
2310 is presented in which the configuration changes
proposal of log database 3 are automatically generated
on occasion of the request from the operations manager
to present the configuration changes proposal to the
operations manager, but the present invention can be
2315 executed even by a method without the aid of manpower
by the operations manager. For example, on occasion
of the management state on log database 3 exceeding a
predetermined threshold in log configuration
management server 8 (examples: a predetermined period
2320 elapsing after the configuration changes, or a data
conservation amount reaching a predetermined level or
more), the configuration changes proposal generation
may be automatically performed and a process up to
automatic reflection of the configuration may be
2325 executed.
94
^^ Thus, the changes proposal of configuration
information including time information is
automatically generated for conservation of timeseries
data for every monitoring target by assisting
2330 the configuration changes work using the present
Example. Therefore, waste of a resource of a timeseries
database of the monitoring platform can be
reduced, and cost required for operations manager
monitoring can be reduced.
2335 [Embodiment 2]
Next, a second embodiment according to the
present invention is described.
The second embodiment refers to an example in
which the configuration changes proposal by the method
2340 described in the first embodiment are once presented
to the operations manager, and then the configuration
changes proposal are readjusted by inputting selection
information on approval or no approval of the
configuration changes by the operations manager.
2345 Figure 21 schematically shows a computer system
assumed in a second embodiment. A constitution of the
computer system is generally similar to the
constitution in the first embodiment. However, the
constitution is different in addition of selection
2350 information T1400 on management database 4, and
95
addition of selection information analysis unit 88 on
log configuration management server 8. Subsequently,
only a different part is described, and the
description on a part similar to the first embodiment
2355 is omitted.
Figure 22 shows one example of selection
information T1400. Selection information T1400
includes data in which a result is recorded for an
operations manager to select approval or rejection to
2360 a recommended screen display of configuration changes
proposal as displayed by configuration changes
proposal displaying unit 87. Column T1401 includes ID
for uniquely identifying configuration of log database
3. The column is associated with column TllOl of
2365 difference information TllOO of the configuration
changes. Column T1402 presents ID for uniquely
specifying individual policy information T900 that the
configuration meets. The column is associated with
column T901 of individual policy information T900.
2370 Here, into the individual policy information that the
configuration meets, a processing result in S215
during data configuration information adjustment
processing, and a processing result in S409 during
index configuration information adjustment processing
2375 are to be registered. Column T1403 includes a result
96
^^ of the operations manager selecting approval or
rejection of change to the configuration. "Approval"
when the change is approved, and "rejection" when the
change is rejected are registered thereinto,
2380 respectively.
Figure 23 shows a sequence of readjustment of
configuration changes proposal in a second embodiment.
First, as a premise, configuration changes proposal
generation processing is performed in a manner similar
2385 to the processing in the first embodiment. Further,
the sequence is in a state in which recommended screen
display 870 of generated configuration changes
proposal is presented to an operations manager. In
the state, the operations manager selects approval or
2390 rejection of each configuration changes on recommended
screen display 870, and presses down resubmitted
proposal button 875, and thus a configuration changes
review request is made. If the configuration review
request is made, selection information analysis unit
2395 88 executes policy automatic adjustment processing S5
based on selection information T1400. Then,
configuration changes proposal generating unit 82
reruns configuration changes proposal generation
processing using adjusted policy information in a
2400 manner similar to the processing in the first
97
^^ embodiment
Figure 24 shows a flowchart of automatic
adjustment processing of policy information of
selection information analysis unit 88 in a second
2405 embodiment. As one example of processing, the present
embodiment presents an example for adjusting
individual policy information T900. Adjusted
individual policy information T900 may be temporarily
used only during a readjustment request or may also be
2410 continuously used thereafter. In the example, the
information is to be temporarily used only during the
readjustment request. First, selection information
analysis unit 88 acquires, upon configuration review
request, individual policy information T900 of the
2415 tenant (S501). In the example, adjusted individual
policy information T900 is, on the assumption that the
information is temporarily used only during the
readjustment request, to copy individual policy
information T900 before adjustment, and retracted.
2420 Next, selection information analysis unit 88 acquires
difference information TllOO of the configuration
changes and selection information T1400 thereof of the
tenant who requires for the change (S502).
Furthermore, the unit 88 enumerates a pair of
2425 difference information TllOO of the configuration
98
changes and selection information T1400 for
configuration subjected to "rejection" among selection
information T1400 (S503).
Subsequently, selection information analysis unit
2430 88 repeats processing in S504 to S506 for the pair
enumerated in S503.
Selection information analysis unit 88 selects an
untested pair PAIR from the pair enumerated in S503
(S504). Next, selection information analysis unit 88
2435 generates a new individual policy using an individual
metrics value of difference information TllOO of PAIR
(S505), and overwrites individual policy information
T900 of a tenant by the new policy (S506). As a
processing method in S505 to S506, for example, a
2440 method is applied in which a metrics value included in
rejected difference information TllOO is compared with
a value of individual policy information T900 that the
configuration meets, and a severer value is used as a
new value. Now, configuration T1152 of difference
2445 information TllOO exemplified in Figure 16 is assumed
to be "rejected." Moreover, the configuration T1152
is assumed to meet policy T951 of individual policy
information T900 exemplified in Figure 10. In the
present case, based on rejected configuration T1152,
2450 the unit 88 updates "time required for configuration
^ ^
99
changes" from "less than 60 minutes" to "less than 50
minutes" for the policy, "criterial response time
change with configuration changes" from "1 second or
more decrease" to "12 seconds or more decrease" and
2455 "data volume change with configuration changes" from
"unlimited" to "less than 18 GB in an increase." If a
plurality of configurations meeting identical
individual policy information are rejected,
configuration is updated on severer conditions by
2460 repetition processing. However, the description is
taken as one example, and therefore another processing
method may be applied. For example, a method may be
applied in which loosest conditions are applied among
a plurality of configurations values that meet
2465 identical individual policy information and are
rejected, or a method may also applied in which an
average of the values is used.
Moreover, the unit 88 records only data in latest
configuration changes in selection information T1400
2470 in the present example. However, a method may also be
applied in which selection information T1400 of plural
times of past configuration changes is recorded to
extract only rejected configuration information
therefrom, and an average of the configuration values
2475 is used.
100
As described above, configuration changes
proposal generating unit 82 generates and presents
configuration changes proposal as in Embodiment 1, and
re-recommends configuration changes proposal using
2480 selection information T1400 to the proposal. Thus,
the operations manager can try any other
recommendation method, when the recommendation made by
configuration changes proposal generating unit 82 is
not useful. On the occasion, configuration changes
2485 proposal in conformity with a current request from the
operations manager can be recommended by giving
selection information T1400 as an input. The process
is effective in allowing further reduction of a manhour
of log database configuration examination work by
2490 the operations manager.
[Embodiment 3]
Next, the third embodiment according to the
present invention is described.
The third embodiment refers to an example for
2495 performing weighing in accordance with an importance
degree of service to be a monitoring target during
configuration changes proposal generation of log
database 3 in the first embodiment.
Figure 25 schematically shows a computer system
2500 assumed in a third embodiment. A constitution of the
101
^ ^ computer system is generally similar to the
constitution in the first embodiment. However, the
constitution is different in addition of service
information T1500 and fault information T1600 on
2505 management database 4, and addition of importance
degree analysis unit 89 on log configuration
management server 8. Subsequently, only a different
part is described, and the description on a part
similar to the first embodiment is omitted.
2510 Figure 26 shows one example of service
information T1500. Service information T1500 includes
information on monitoring of each service of each
tenant, and data presenting a role and an importance
degree of service. Service information T1500 is
2515 assumed to be preliminarily input. Column T1501
includes a tenant name managing service. Column T1502
includes a name for uniquely specifying service. For
example, names such as "service 1" and "service 1
demonstration" are registered. Columns T1503 and
2520 T1504 include information on monitoring of each
service of each tenant. Column T1503 includes a list
of data type names for monitoring the service. In the
column, all of data type names for the service are
enumerated, for example, "Web event log of service 1,
2525 CPU utilization rate of OS, ...." Each data type name
102
^^ in the column is corresponded to a data type name on
any other table, such as column T104 of raw data TIOO.
Column T1504 includes a list of instance names for
provision of the service. In the column, all of
2530 instance names for provision of the service are
enumerated, for example, "aplserverl, aplserver2,
apldbl ...." Each instance name in the column is
corresponded to an instance name on any other table,
such as column T105 of raw data TIOO. Columns T1505
2535 and T1506 include data presenting a role and an
importance degree of service. Column 1505 presents
the role of service. For example, data presenting the
role of the service, such as "Web Service production
environment" and "trial environment" are registered.
2540 Column T1505 includes the importance degree of the
service. An importance degree according to the role
of the service is registered into the importance
degree. In the present example, the importance degree
is expressed as a numeric value between "0 and 1."
2545 Moreover, the importance degree is to be higher as the
value is larger. In the present example, the
importance degree of service is preliminarily given,
but may be determined by calculation from any other
information. For example, the importance degree may
2550 be determined from a calculation expression using an
103
0^ end user count using the service as an input.
Moreover, in the present example, the role and the
importance degree are determined for every service,
but may be determined for every instance. For
2555 example, "AP server" or "DB server" may be registered
as the role of the instance, and the importance degree
may be determined according to the value.
Figure 27 shows one example of fault information
T1600. Fault information T1600 includes data for
2560 presenting a record of fault occurrence of each
service. Columns T1601 and T1602 include a tenant
name and a service name, respectively. The columns
are associated with columns T1501 and T1502 of service
information T1500. Column T1603 includes fault
2565 occurred time. In the column, time on which a fault
of service is generated is recorded. Column T1604
includes a data type name used for root cause
analysis. Moreover, column T1605 includes an instance
name causing the fault. Data in columns T1604 and
2570 T1605 are associated with a data type name and an
instance name, such as columns T103 and T104 of raw
data TlOO, respectively.
Figure 28 shows a sequence of configuration
changes proposal generation in the third embodiment.
2575 Importance degree analysis unit 89 calculates, upon
104
configuration changes proposal review request from an
operations manager, an importance degree of service
using service information T1500 or fault information
T1600 or both thereof (S6) . Next, the unit 89
2580 executes configuration changes proposal generation
processing similar to the processing in Embodiment 1.
However, a value weighed according to the importance
degree is used for a configuration value of retention
on configuration information . Moreover, the value
2585 weighed according to the importance degree is also
used for individual policy information T900 in a
similar manner.
Hereinafter, an example of an analytical method
for the importance degree by importance degree
2590 analysis unit 89 is given. Here, an example for
calculating the importance degree for every monitoring
target is presented as the importance degree of
service. For example, as a method for calculating the
importance degree from service information T1500, a
2595 method is conceivable in which all rows including the
monitoring target are extracted from service
information T1500 to calculate an average of the
importance degree T1506, and the value is determined
as the importance degree of the monitoring target.
2600 Moreover, as a method for calculating the importance
105
^^ degree from fault information T1600, for example, a
method is conceivable in which occurrence frequency of
a fault during a review target period is tabulated for
every monitoring target to treat height of the
2605 frequency as the importance degree.
Importance degree analysis unit 89 calculates the
importance degree for every monitoring target, and
then configuration changes proposal generating unit 82
performs configuration information generation
2610 processing in consideration of the importance degree.
The fundamental processing content is similar to the
content in Embodiment 1. However, the content is
different in weighing according to the importance
degree a configuration value of configuration
2615 information and each criterial value of policy
information as generated. For example, in processing
S108 and S313 in which retention is configured,
calculation is made by multiplying a log utilization
day count of the retention by a weighing factor that
2620 becomes longer in the period accordingly as the
importance degree of the monitoring target is higher
(example: retention = log utilization day count x
(importance degree + 1)). In a similar manner,
calculation is also made by multiplying individual
2625 policy information T900 by the weighing factor
106
2630
according to the importance degree of the monitoring
target (example: data volume conditions are further
relieved accordingly as the importance degree is
higher).
As described above, configuration changes
proposal generating unit 82 performs weighing
according to the importance degree to generate
configuration changes proposal, thereby allowing
conservation of more important log data for a longer
2635 period in comparison with the case in Embodiment 1.
Therefore, loss caused by excessive log data deletion
(fault correspondence man-hour increase, opportunity
loss) that has been so far generated can be further
suppressed in more important service and log data.
2640 Accordingly, the present invention is effective in
allowing further intensification of log monitoring of
the monitoring platform by the operations manager.
In the first to the second embodiments as
described above, the description is made for
2645 configuration changes proposal generation processing
regarding raw data configuration information,
compressed data configuration information and index
configuration information. However, the embodiment
according to the present invention is not limited to
2650 the configuration information generation. Therefore,
107
2655
2660
the present invention may be applied to configuration
information generation of other than the description
above. For example, when configuration for caching a
result obtained by analyzing a plurality of log data
exists, the present invention may also be applied to
configuration changes proposal generation of the
configuration information.
As described above, the embodiments according to
the present invention are described in full detail
with reference to drawings, but a specific
constitution is not limited to the embodiments, and
includes a design or the like in the range without
departing from the scope of the present invention.

10
2670
2675
••t ' ^i ^7 (
What is claimed is: r^^f '"• ^^
1. A time-series database configuration automatic
generation system comprising
a time-series database and
a monitoring server using the time-series database,
characterized in that,
the monitoring server has
a function for storing and searching time-series data
using the time-series database, respectively,
regarding a plurality of monitoring targets,
a configuration management function for managing
configuration information regarding a kind and a
conservation format of the time-series data for every
monitoring target, and
2680 a function for generating changes proposal of the
configuration information including time information
regarding each monitoring target, and
the monitoring target comprises a combination of a
tenant being a service provision target, a data type
2685 and an instance name for uniquely identifying an
instance name of equipment of the tenant, and
changes proposal of the configuration information
including the time information is generated for every
monitoring target based on current configuration
2690 information on the time-series database, an operation
losD^
- , ^ ^'
%3 ' •' 1^
^>k history of the time-series data and information on an
amount of the time-series data.
2. The time-series database configuration
2695 automatic generation system, characterized in that,
in claim 1,
the monitoring server has
a function for generating a metrics in which an
influence and an effect in connection with change of
2700 the configuration are quantified, and
a function for presenting the metrics together with
the configuration changes proposal to a terminal of an
operations manager of the tenant for every monitoring
target.
2705
3. The time-series database configuration
automatic generation system, characterized in that,
in claim 2,
the monitoring server
2710 holds policy information in which conditions to be met
by the metrics is preliminarily described as a level
value as a request for change of the configuration
information for every monitoring target, and
has a function for comparing the level value described
2715 in the policy information with a metrics value of the
r.fit's
^ changes proposal of the configuration information.
judging whether the configuration changes proposal
meet the request, and generating the configuration
changes proposal based on a result of the judgment,
2720 a function for seeking judgment of the operations
manager of the tenant for the configuration changes
proposal displayed on the terminal, and
a function for registering as new configuration
information the configuration changes proposal
2725 approved in the judgment.
4. The time-series database configuration
automatic generation system, characterized by,
in claim 3,
2730 having
a function for recording as selection information a
result that the operations manager selects approval or
rejection to the configuration changes proposal,
a function for analysis the request for the
2735 configuration changes of the operations manager using
the selection information of the configuration
changes, and updating the policy information based on
the result of analysis, and
a function for generating the configuration changes
2740 proposal using the updated policy information.
^
2745
Yll ^^r^ r^ V Vi ?- ^,
v_>
l . ^ V i^
5. The time-series database configuration
automatic generation system, characterized by,
in claim 3,
having
a function for calculating an importance degree of the
monitoring target in advance, and
a function for weighing a configuration value of the
configuration information of the time-series database
2750 or the policy information according to the calculated
importance degree.
6. The time-series database configuration
automatic generation system, characterized in that,
2755 in claim 5,
the monitoring server
holds information on monitoring of each service of
each tenant, and service information in which data
presenting a role and an importance degree of the
2760 service are preliminarily described, for every
monitoring target, and
has a function for calculating the importance degree
using the service information.
2765 The t i m e - s e r i e s database c o n f i g u r a t i on
automatic generation system, characterized m that,
in claim 5,
the monitoring server
holds fault information in which fault occurrence is
2770 recorded to each service for every monitoring target
and
has a function for calculating the importance degree
using the fault information.
2775 8. The t i m e - s e r i e s database c o n f i g u r a t i on
automatic generation system, characterized in that,
in claim 2,
the metrics includes
a first metrics value for evaluating time required for
2780 the configuration changes,
a second metrics value for evaluating change of
performance in connection with the configuration
changes, and
a third metrics value for evaluating change of a data
2785 volume in connection with the configuration changes,
and
the monitoring server has
a function for generating difference information
including the first to third metrics values in
2790 connection with the configuration changes, and
(^% information on an increase and decrease of a charged
amount to the tenant, corresponding to each difference
information, and displaying a result thereof together
with the configuration changes proposal on a terminal
2795 of the operations manager of the tenant.
9. The time-series database configuration
automatic generation system, characterized by,
in claim 1,
2800 comprising a configuration changes proposal generating
unit for the generating configuration changes proposal
in conformity with data actually accessed the tenant,
in which the configuration changes proposal generating
unit includes
2805 a data configuration information generating unit for
acquiring, upon configuration review request from the
tenant, an operation history during a designated
review target period of the tenant, and generating
data configuration information,
2810 a data configuration information adjusting unit for
judging and adjusting whether the generated
configuration information meets a preset policy of
configuration changes and generating the configuration
changes proposal,
2815 an index configuration information generating unit for
1>
114 " n
# ^ p r o v i d i n g t h e t i m e - s e r i e s d a t a w i t h an'^index for .,p>j
ftC^
accelerating data search,
an index configuration information adjusting unit for
judging whether the configuration information meets
2820 the preset policy of the configuration changes, and
adjusting the configuration information, and
a configuration changes proposal displaying unit for
displaying a proposal of the configuration changes on
a screen of the terminal of the manager of the tenant.
2825
10. A time-series database configuration automatic
generation method in a data search system comprising a
time-series database and a monitoring server using the
same, characterized in that,
2830 the monitoring server comprises a configuration
management function for managing configuration
information regarding conservation of the time-series
data, respectively, regarding a plurality of
monitoring targets, and
2835 the monitoring target comprises a combination of a
tenant being a service provision target, a data type,
and an instance name for uniquely identifying
equipment of the tenant, and
time-series data using the time-series database are
2840 stored and searched for every monitoring target.
^
•115 o J
^^1 V
a changes proposal of the configuration information
including time information is generated for every
monitoring target, and
a changes proposal of the configuration information
2845 including the time information is generated for every
monitoring target based on current configuration
information of the time-series database, an operation
history of the time-series data and information on an
amount of the time-series data.
2850
11. The time-series database configuration
automatic generation method, characterized by,
in claim 10,
generating a metrics in which an influence and an
2855 effect in connection with change of the configuration
is quantified for every monitoring target, and
presenting the metrics together with the configuration
changes proposal to a terminal of an operations
manager of the tenant.
2860
12. The time-series database configuration
automatic generation method, characterized by,
in claim 11,
generating a metrics in which an influence in
2865 connection with change of the configuration is
2880
2885
quantified using a changes proposal of the -^.{V] 7ut
configuration information for every monitoring target,
displaying the metrics together with the configuration
changes proposal on the terminal of the operations
2870 manager of the tenant, seeking for judgment of
approval or rejection to the changes proposal, and
registering as new configuration information the
configuration changes proposal approved in the
j udgment.
2875
13. The time-series database configuration
automatic generation method, characterized by,
in claim 12,
displaying on the terminal of the operations manager
of the tenant difference information of the metrics in
connection with the configuration changes, and an
increase and decrease amount of a charged amount to a
manager of the tenant, corresponding to the difference
information.
14. A monitoring server using a time-series
database, the monitoring server of the time-series
database being characterized in that,
the time-series database has
2890 a function for holding time-series data of a tenant
^fe being a service provision target, ^T^, - ' w-' nf\\'X
the monitoring server has
a function for storing and searching time-series data
using the time-series database, respectively,
2895 regarding a plurality of monitoring targets,
a configuration management function for managing
configuration information regarding a kind and a
conservation format of the time-series data for every
monitoring target, and
2900 a function for generating a changes proposal of the
configuration information including time information
regarding each monitoring target,
the monitoring target comprises a combination of the
tenant, a data type, and instance name for uniquely
2905 identifying equipment of the tenant, and
a changes proposal of configuration information
including the time information is generated for every
monitoring target based on current configuration
information of the time-series database, an operation
2910 history of the time-series data and information on an
amount of the time-series data.
15. The monitoring server of the time-series
database, characterized by,
2915 in claim 14,
118,...-•'•r 'l^
^ ^ having a function for generating a metrics in which an
influence and an effect in connection with change of
the configuration are quantified, and
a function for generating difference information of
2920 the metrics value in connection with change of the
configuration, and information on an increase and
decrease of a charged amount to the tenant,
corresponding to each difference information.
16. A time-series database configuration
2925 automatic generation system, substantially as herein
described with reference to accompanying drawings and
examples.
17. A time-series database configuration
2930 automatic generation method, substantially as herein
described with reference to accompanying drawings and
examples.
18. A monitoring server using a time-series
2935 database, substantially as herein described with
reference to accompanying drawings and e }j-*fl(ip 1 e s .