9 Siebel Metrics

This chapter provides descriptions for all Siebel metric categories, and provides tables that list and describe associated metrics for each category.

Siebel metrics consist of the following categories:

Siebel Component Metrics

Table 9–1 provides details about the Siebel component metrics.

Table 9-1 Siebel Component Metrics

Metric Description and User Action Data Source

Average Object Manager Response Time (in milliseconds)

This metric shows the average (mean) time required to process a request sent to the Object Manager. It corresponds to the responsiveness of the interactive user sessions. This metric is primarily useful for establishing long-term performance trends and capacity planning. If application performance is deteriorating, the metric value rises.

Since response time can be affected by factors such as the efficiency of the code, efficiency of your application configuration, server CPU and memory capacity, end-user behavior and data volume, you need to first identify the root cause of slower response time before taking any action. See the Siebel Performance Tuning Guide for more information.

Average Object Manager Response Time = Total Object Manager Response Time of all requests on all Object Manager sessions / Total number of requests of all Object Manager sessions

Average Connect Time for Object Manager Sessions (in seconds)

In establishing an interactive session with Object Manager, Object Manager needs to perform many tasks such as authentication, initialization, and allocating the necessary resources. The time required to perform all these activities is the connect time. The Average Connect Time for Object Manager sessions is the average amount of time required to establish a connection to a particular Object Manager since the startup of the component. This metric is primarily useful for evaluating connection performance overtime.

Average Connect Time = Total Connect Time / Total Number of Connections

Average Number of Requests Per Object Manager Session

Users can perform multiple actions, such as querying records, updating records, and clicking a button to issue a command in a single session. Each of these actions corresponds to one or more requests sent from the user's browser to the Object Manager. The Average Number of Requests for each Object Manager session is the average (mean) number of requests sent to Object Manager in all the Object Manager sessions recorded after the Siebel Enterprise startup.

This metric is intended for informational purposes only. You can use it to track the usage pattern of your users to determine how much processing load they generate in a session. The information is especially useful over a period of time. When you combine the long-term trending of this metric with Total Object Manager sessions, you can determine whether the processing load is increasing or decreasing over time. The information can then be used in making capacity planning decisions.

Average Number of Requests per Object Manager = Total Number of Requests of all Object Manager sessions / Number of Object Manager sessions

Average Size of Reply Messages (in bytes)

This metric shows the size of the response to user-submitted requests to Object Manager. A greater size indicates more data is being passed.

This metric is intended primarily for establishing long-term performance trends. If the metric trends up, it shows more data is being passed, which can be caused by factors such as increasing data volume, users issuing queries that return more data, and application configuration changes. You must identify the root cause to address the trend. Refer to the Siebel Performance Tuning Guide for more information.

This metric is computed by dividing the Total Size of all reply messages from all requests after the startup of the Object Manager by the number of requests.

Average Size of Request Messages (in bytes)

This metric shows the size of the request to user-submitted requests to Object Manager. A greater size indicates more data is being passed in submitting the request.

This metric is intended primarily for establishing long-term performance trends. If the metric trends up, it shows that more data is being passed, which can be caused by factors such as application configuration changes. You must identify the root cause to address the trend. Refer to the Siebel Performance Tuning Guide for more information.

This metric is computed by dividing the Total Size of all request messages from all requests since the startup of the Object Manager by the number of requests.

CPU Usage

This metric shows the amount of CPU time consumed by this Siebel component. CPU data originates from two different sources: OS level (process-based statistics gathered by the Oracle Agent) and Srvrmgr statistics.

This metric is intended primarily for establishing long-term performance trends. If the value of this metric increases over time, it shows that more intensive processing is occurring on the application server for this component. The change may be caused by application configuration changes or changes to the underlying Siebel software. You need to identify the root cause to address the trend. Refer to the Siebel Performance Tuning Guide for more information.

This metric is calculated by adding the CPU time of all component tasks for the component after the Siebel Server start up.

Max. MTS

This metric sets the maximum number of multi-threaded shells. You should review the Running MTS Processes metric periodically to learn about the level of running processes. If the number of MTS running is often close to exceeding the limit set through Max MTS, consider increasing the Max MTS value. On the other hand, if the number of MTS is always a fraction of the limit, consider lowering Max MTS.

The metric is set on the Siebel Component.

Number Component Process Failures

This metric provides the count of component tasks that exited with errors. Component Tasks exit with errors for many reasons, ranging from not having the correct business data to work with to failure in the software. If the number of such failures increases dramatically, you must examine what is wrong. Start by examining the Alert log and the Siebel Server Manager to find out the Tasks that exited with errors.

Component Tasks that exited with errors are counted.

Run State

This metric shows the current status of the component. If a component is down or disabled when it is not supposed to be, you should try to restart it or enable it using Siebel Server Manager. Also examine the component log file for information on why the component is not working.

The value of this metric is updated at each sampling period.

Running MTS Processes

This metric shows the number of running multi-threaded shell (MTS) processes. This metric is intended for informational purposes and can be used for long-term trending to analyze the component usage. If the number of MTS processes trends upward, it indicates increased usage of the component. Tuning or capacity adjustment may be required.

The value of this metric is queried from the Siebel Server at each sampling period.

Start Time

This metric is intended for informational purposes, and can be used during diagnostics to determine when the component was started.

The start time of the component. This metric is collected from the Siebel Server during each sampling period. It typically does not change unless the component is restarted.

Current Tasks

This metric shows the current number of running tasks for the component. Since the maximum number of tasks is limited by MaxTasks, this parameter indirectly indicates the number of remaining tasks. If the number of tasks becomes too high relative to MaxTasks, you may need to determine if there are runaway tasks or whether MaxTasks is set too low.

The value of Current Tasks is updated for each sampling period by querying the Siebel Server.

Max Tasks

This metric determines the maximum number of tasks that can be run for a Siebel Server Component. If the current number of running tasks equals Max Tasks, no new tasks can be spawned. For Interactive Object Managers, it means that no new users can log on. It is very important to make sure that Max Tasks is set to a level higher than the expected maximum number of tasks to be run at a given time. Refer to the Siebel System Administration Guide for more information.

You should review the Current Tasks metric periodically to learn about the level of task usage. If the number of tasks running is often close to exceeding the limit set by Max Tasks, consider increasing the Max Tasks value. On the other hand, if the number of tasks is always a fraction of the limit, consider lowering Max Tasks.

This mertric is set using Siebel Server Manager.

Average Time for SQL Execute Operations (in seconds)

This metric shows the average (mean) amount of time the database requires to process the SQL statement after the statement is parsed. The metric is only applicable to older Siebel components such as EIM, and does not show the time required to process Object Manager requests.

This metric is intended primarily for establishing long-term performance trends. If the value of this metric increases over time, it shows that SQL statements involving more complex operations are being executed, or the database has become less efficient at processing requests. These can be caused by changes to the application, increasing data volume, or tuning changes to the database. You must identify the root cause to address the trend. Refer to the Siebel Performance Tuning Guide for more information.

This metric is computed by dividing the total time for SQL execution from all requests since the startup of the component by the number of requests.

Average Time for SQL Parse Operations (in seconds)

This metric shows the average (mean) amount of time the database requires to parse the SQL statements being passed to it. The metric is only applicable to older Siebel components such as EIM, and does not show the time required to process Object Manager requests.

This metric is intended primarily for establishing long-term performance trends. If the value of this metric increases over time, it shows that SQL statements have become more complex. Since Siebel generates all SQL statements dynamically, the change could be caused by changes to the underlying Siebel product, to the complexity of the business components, or query specifications you defined. You must identify the root cause to address the trend. Refer to the Siebel Performance Tuning Guide for more information.

This metric is computed by dividing the total time for parsing SQL statements from all requests for this component since the startup of the Siebel Server by the number of requests.

Total Number of SQL Execute Operations

This metric shows the total number of SQL statements executed. The number is aggregated from completed component tasks; that is, it does not reflect SQL statements executed by currently active sessions.

This metric is intended primarily for establishing long-term performance trends. If the value of this metric increases over time, it shows increased access to the database, which may necessitate tuning or capacity adjustments.

Total Number of SQL Fetch Operations

This metric shows the total number of SQL fetch operations for the component.

This metric is queried from the Siebel Server at each sampling period.

Total Number of SQL Parse Operations

This metric shows the total number of SQL statements parsed. The metric is only applicable to older Siebel components such as EIM, and does not show the time required to process Object Manager requests. It is used in computing the Average time for SQL Parse operations.

This metric is intended primarily for establishing long-term performance trends. If the value of this metric increases over time, it shows increased access to the database, which may necessitate tuning or capacity adjustment.

This metric is queried from the Siebel Server at each sampling period.

Average CPU Time

Average CPU Time shows the average (mean) amount of time consumed to process a request. A higher CPU Time indicates more CPU-intensive processing.

This metric is primarily useful in establishing long-term performance trends and capacity planning. If the metric trend is up, this indicates more CPU-intensive processing. You may need to add more processing capacity or tune the application configuration if this situation continues.

This metric is computed by dividing the Total CPU Time of all requests after the component startup by the Total Number of Requests after the component startup.Average CPU Time = Total CPU Time / Number of Requests.

Average Response Time

The Average Response Time shows the average (mean) time required to respond to an end-user action. It corresponds to the responsiveness of the interactive user sessions during a measurement period. This metric is useful for short term monitoring as well establishing long-term performance trends. If application performance is deteriorating, the metric value increases.

Since response time can be affected by factors such as the efficiency of the code, efficiency of your application configuration, server CPU and memory capacity, end-user behavior, and data volume, you need to first identify the root cause of slower response time before taking any action. See the Siebel Performance Tuning Guide for more information.

This metric is computed via SARM data, as opposed to Server Component metrics.Average Response Time = Total Response Time all requests / Total number of requests during a measurement period.

Memory Usage

Memory usage measures the total amount of memory consumed by the running tasks of the component. This metric is intended primarily for informational purposes, especially in diagnosing memory-related problems. Constantly increasing memory usage may indicate a memory leak.

Memory consumption for all component processes is retrieved from the operating system.

Number of Requests

This metric indicates the total number of requests submitted to the component since the start-up of the Siebel Server. This metric is used as the denominator in calculating many other summary metrics.

The value is intended for informational purposes and can be used for long-term trending to analyze the usage of the component. If the number of requests trends up, it indicates increased usage of the component. Tuning or capacity adjustment may be required.

The value of Number of Requests is updated for each sampling period by querying the Siebel Server.

Status

This metric shows the current status of the component. It is the binary representation of the component's availability: 1 = Available, and 0 = Down. If a component is down or disabled when it is not supposed to be, you need to attempt to restart it or enable it using Siebel Server Manager. Also, you can review the component log file for information on why the component is not working.

The value of this metric is updated at each sampling period.

Total CPU Time

This metric shows the aggregated CPU time used by all tasks completed for this component since the component was started. This metric is intended only for informational use. In most cases, the current CPU consumption and trends in CPU consumption are more informative. These can be checked with the CPU Usage metric.

This metric is collected from the Siebel server.

Total Response Time

This metric shows the total response time for the component.

This metric is derived from the srvrmgr statistics.


Siebel Component Group Target Metrics

Table 9–2 provides details about the Siebel component group target metrics.

Table 9-2 Siebel Component Group Target Metrics

Metric Description and User Action Data Source

Start Time

This metric shows the start time of the component. This metric is intended for informational purposes, and can be used during diagnostics to determine when the component was started.

This metric is collected from the Siebel Server during each sampling period. It typically does not change unless the component is restarted.

Status

This metric shows the current status of the component. It is the string representation of the state, such as Running, Online, and Shutdown. If a component is down or disabled when it is not supposed to be, you need to attempt to restart it or enable it using Siebel Server Manager. Also, you can review the component log file for information on why the component is not working.

The value of this metric is updated at each sampling period, and it is retrieved using srvrmgr.


Siebel Gateway Target Metrics

Table 9–3 provides details about the Siebel Gateway target metrics.

Table 9-3 Siebel Gateway Target Metrics

Metric Description and User Action Data Source

Response

This metric shows the response time of the Siebel Gateway.

Retrieved using srvrmgr.

Status

This metric shows the current status of the Siebel Gateway target.

Retrieved using srvrmgr.


Siebel Server Target Metrics

Table 9–4 provides details about the Siebel Server target metrics.

Table 9-4 Siebel Server Target Metrics

Metric Description and User Action Data Source

File System Usage (%)

This metric shows the percentage of disk space used for the selected file system. The file systems covered are the Siebel installation directory, Siebel log directory, and Siebel file system used to share documents across Siebel servers. If one of the file systems is close to being exhausted (for example, 95% fill level), the administrator should consider cleaning up the respective file system or adding additional disk space.

File system monitoring

File System Usage (KB)

This metric shows the absolute amount of disk space used for the selected file system. The file systems covered are the Siebel installation directory, Siebel log directory, and Siebel file system used to share documents across Siebel servers.

File system monitoring

Average Connect Time for Object Manager Sessions (in seconds)

In establishing an interactive session with Object Manager, it needs to perform many tasks such as authentication, initialization, and allocating the necessary resources. The time the Object Manager requires to perform all these activities is the connect time. The Average Connect Time for Object Manager sessions is the average amount of time necessary to establish a connection to a particular Object Manager after starting the component. Metrics from all active object managers running inside a Siebel server are aggregated to the Siebel server level to provide this metric. This metric is primarily useful for evaluating connection performance overtime.

Average Connect Time = Total Connect Time / Total Number of Connections

Average Number of Requests Per Object Manager Session

Users can perform multiple actions, such as querying records, updating records, and clicking a button to issue a command in a single session. Each of these actions corresponds to one or more requests sent from the user's browser to the Object Manager. The Average Number of Requests for each Object Manager session is the average (mean) number of requests sent to the Object Manager in all the Object Manager sessions recorded after starting the Siebel Enterprise. Metrics from all active object managers running inside a Siebel server are aggregated to the Siebel server level to provide this metric.

This metric is only intended for informational purposes. You can use it to track the usage pattern of your users to see how much processing load they generate in a session. The information is especially useful over time. When you combine the long-term trending of this metric with the Total Object Manager session, you can see whether the processing load is increasing or decreasing over time. The information can then be used in making capacity planning decisions.

Average Number of Requests per Object Manager = Total Number of Requests of all Object Manager sessions / Number of Object Manager sessions

Average Object Manager Response Time (in milliseconds)

This metric shows the average (mean) time necessary to process a request sent to the Object Manager. It corresponds to the responsiveness of the interactive user sessions. Metrics from all active object managers running inside a Siebel server are aggregated to the Siebel server level to provide this metric.

This metric is primarily useful for establishing long-term performance trends and capacity planning. If application performance is deteriorating, the metric value increases. Since response time can be affected by factors such as efficiency of code, efficiency of your application configuration, server CPU and memory capacity, end-user behavior and data volume, you first need to identify the root cause for slower response time before taking any action. See the Siebel Performance Tuning Guide for more information.

Average Object Manager Response Time = Total Object Manager Response Time of all requests on all Object Manager sessions of this server / Total number of Requests of all Object Manager sessions

Average Size of Reply Messages (in bytes)

This metric shows the size of the response sent from user-submitted requests to Object Manager. A large size indicates that a large amount is being passed. Metrics from all active object managers running inside a Siebel server are aggregated to the Siebel server level to provide this metric.

This metric is intended for establishing long-term performance trends. If this metric trends upward, it shows that a large amount of data is being passed, which can be caused by factors such as increasing data volume, users issuing queries that return more data, and application configuration changes. You need to identify the root cause of the problem to address this trend.

This metric is computed by dividing the Total Size of all reply messages from all requests since the startup of the server by the number of requests.

Average Size of Request Messages (in bytes)

This metric shows the size of the request to user-submitted requests to Object Manager. A greater size indicates that more data is being passed while submitting the request. Metrics from all active object managers running inside a Siebel server are aggregated to the Siebel server level to provide this metric.

This metric is intended primarily for establishing long-term performance trends. If the metric trends upward, it shows that more data is being passed, which can be caused by factors such as application configuration changes. You need to identify the root cause to address the trend. Refer to the Siebel Performance Tuning Guide for more information.

This metric is computed by dividing the Total Size of all request messages of all requests since the startup of the server by the number of requests.

Average Time for SQL Execute Operations (in seconds)

This metric shows the average (mean) amount of time the database requires to process the SQL statement after the statement is parsed. The metric is only applicable to older Siebel components such as EIM, and does not show the time required to process Object Manager requests. Metrics from all active object managers running inside a Siebel server are aggregated to the Siebel server level to provide this metric.

This metric is intended primarily for establishing long-term performance trends. If the value of this metric increases over time, this shows that SQL statements involving more complex operations are being executed, or the database has become less efficient at processing requests. These can be caused by changes to the application, increasing data volume, or tuning changes to the database. You need to identify the root cause to address the trend. Refer to the Siebel Performance Tuning Guide for more information.

This metric is computed by dividing the total time for SQL execution from all requests after starting the server by the number of requests.

Average Time for SQL Fetch Operations (in seconds)

This metric shows the average (mean) amount of time the database requires to fetch records by an SQL statement after the statement is parsed. The metric is only applicable to older Siebel components such as EIM, and does not show the time required to process Object Manager requests. Metrics from all active object managers running inside a Siebel server are aggregated to the Siebel server level to provide this metric.

This metric is intended primarily for establishing long-term performance trends. If the value of this metric increases over time, this shows that SQL statements involving more complex operations are being executed, or the database has become less efficient at processing requests. These can be caused by changes to the application, increasing data volume, or tuning changes to the database. You need to identify the root cause to address the trend. Refer to the Siebel Performance Tuning Guide for more information.

This metric is computed by dividing the total time for SQL execution from all requests since the startup of the server by the number of requests.

Average Time for SQL Parse Operations (in seconds)

This metric shows the average (mean) amount of time the database requires to parse the SQL statements being passed to it. The metric is only applicable to older Siebel components such as EIM, and does not show the time required to process Object Manager requests.

This metric is intended primarily for establishing long term performance trends. If the value of this metric increases over time, it shows that SQL statements have become more complex. Since Siebel generates all SQL statements dynamically, the change could be caused by changes to the underlying Siebel product, changes to the complexity of the business components, or query specifications you defined. You need to identify the root cause to address the trend. Refer to the Siebel Performance Tuning Guide for more information.

This metric is computed by dividing the total time for parsing SQL statements from all requests for this component since the startup of the Siebel Server by the number of requests.

CPU Usage

This metric shows the amount of CPU time this Siebel component consumed. This metric is intended primarily for establishing long-term performance trends. If the value of this metric increases over time, it shows that more intensive processing is occurring on the application server for this component. The change may be caused by application configuration changes or changes to the underlying Siebel software. You need to identify the root cause to address the trend. Refer to the Siebel Performance Tuning Guide for more information.

The metric is computed by adding up the CPU time of all component tasks for the component since the startup of the Siebel Server.

File System Free Space (KB)

This metric shows the absolute amount of free disk space currently available in the selected file system. The file systems covered are the Siebel installation directory, the Siebel log directory, and the Siebel file system used to share documents across Siebel servers. If the absolute amount of free disk space becomes very small (for example, less than 500 MB), there is a risk that the file system will fill up quickly, which would affect availability of the Siebel server. If this value becomes too low, the file system must be cleaned, or additional disk space needs to be added.

File system monitoring

Memory Usage

Memory usage measures the total amount of memory consumed by the processes running as part of the Siebel server. The metric is primarily intended for informational purposes, especially in diagnosing memory-related problems. Constantly increasing memory usage may indicate a memory leak.

Memory consumption for all Siebel server processes is retrieved from the operating system.

Number Component Process Failures

This metric provides the count of component tasks that exited with errors. Component Tasks exit with errors for many reasons, ranging from not having the correct business data to work with to failure in the software. If the number of such failures increases dramatically, something is definitely wrong and should be examined. A good place to start would be to examine the Alert log and the Siebel Server Manager to find out which Tasks exited with errors.

Component Tasks that exited with errors are counted.

Number Component Process Restarts

This metric provides the count of component tasks that exited with errors. Component tasks exit with errors for many reasons, ranging from not having the correct business data to work with, to failure in the software. If the number of such failures increases dramatically, something is definitely wrong and should be examined. A good place to start would be to examine the Alert log and the Siebel Server Manager to find out which tasks exited with errors.

Component Tasks that exited with errors are counted.

Number of Retries Due to DB Connection Loss

When the Siebel Server loses connection to the database, it attempts to retry the operation before stopping and reporting the problem. This metric shows the number of retries. If the number of retries rises dramatically, this indicates problems with the database or the network, and the administrator should examine these components to determine why they are dropping database connections.

 

Number of Retries Due to Deadlock Rollbacks

When the Siebel Server loses connection to the database, it attempts to retry the operation before giving up and reporting the problem. This metric records the number of retry attempts. If the number of retries increases dramatically, this indicates problems with the database or the network, and the administrator should examine these components to determine why they are dropping database connections.

 

Number of Times All Retries are Exhausted

This metric counts the number of retry failures; that is, the number of times the Siebel Server stops the attempts of communicating with the database. A database administrator should be called to find out why the Siebel Server cannot communicate with the database.

 

Start Time

This metric shows the start time of the Siebel server.

Retrieved by srvrmgr from the running Siebel server.

Status

This metric shows the current status of the Siebel server (that is, whether it is available or down). If the status of a Siebel server is shown as unavailable, the administrator should check the reason of the failure and attempt to start the Siebel server again.

The status of a Siebel server is determined by running the srvrmgr command line utility.

Available Disk space (KB)

This metric shows the absolute amount of disk space available in the selected file system. The file systems covered are the Siebel installation directory, the Siebel log directory, and the Siebel file system used to share documents across Siebel servers.

File system monitoring

Total CPU Time for Component Tasks (in seconds)

The total CPU time in seconds for component tasks.

 

Total Number of Level 0 and 1 Errors

This metric counts the number of fatal errors and regular errors on the Siebel Server. There are always a small number of errors on the Siebel Server. However, if the count develops an upside trend, the administrators should check the alert and error logs.

 

Log Tracking Metric

This metric is used to send alert notifications to on-call administrators when specific patterns are found in the middleware-related log files.

 

Siebel Workflow Target Metrics

Table 9–5 provides details about the Siebel Workflow target metrics.

Table 9-5 Siebel Workflow Target Metrics

Metric Description

Total Number of Completed Process Instances in Past 1 Hour

This metric shows the total number of workflow process instances that completed in the past hour. The collection frequency is every 15 minutes.

Total Number of Workflow Policy Instances in Waiting State

This metric shows the total number of workflow policy instances waiting in a queue. The collection frequency is every 15 minutes.

Total Number of Workflow Process Instances in Waiting State

This metric shows the total number of workflow process instances waiting in a queue. The collection frequency is every 15 minutes.

Number of Monitored Process Instances Failed State

This metric shows the number of workflow process instances that failed. The collection frequency is every 15 minutes.

Number of Monitored Process Instances in Waiting State

This metric shows the total number of workflow process instances waiting in a queue. The collection frequency is every 15 minutes.