Custom health condition subexpression builder

Use the custom health condition subexpression builder to define a custom health condition for the health policy. Use the build subexpression utility to build complex rule conditions from subexpressions by using AND, OR, NOT and parenthetical grouping. The subexpression builder validates the rule when you apply the changes, and alerts you to mismatched parentheses and unsupported logic operators. To view this console page, click Operational polices > Health policies > New. If we choose a custom health condition, the Run reaction plan when field is displayed. Click Subexpression builder to build the custom health condition.

Select the properties to include in the custom health condition, and click Generate subexpression. The subexpression value displays. To append the subexpression to your custom health condition, click Append.

Logical operator

Specifies the operator used to append this subexpression to the previous subexpression in the custom health condition.

and

Both subexpressions that are around the and operator must be true for actions to be taken on the health policy.

or

To select a node, one of the two subexpressions that are around the or operator must be true for actions to be taken on the health policy.

Operand

• PMI metric: From server start and PMI metric: From last interval

  The PMI metric: From server start operand uses an average number of the reported values from the time that the server started.

  The PMI metric: From last interval operand uses an average of the reported values in the last interval. The interval is the length of the health controller cycle.

  Both PMI operands have the following PMI modules:

  • Connection pool module (JDBC): These metrics can be used only on servers that are running WebSphere Application Server. The connection pool module has the following metrics:

    • Faults: Number of connection timeouts in the pool.

    • Number of creates: Total number of connections created.

    • Percent used: Average percent of the pool that is in use. The value is based on the total number of configured connections in the connection pool, not the current number of connections.

    • Prepared statement cache discards: Specifies the number of statements that are discarded because the cache is full.

    • Number of destroys: Total number of connections that are closed.

    • Pool size: Size of the connection pool.

    • Number of connection handles: Specifies the number of connection objects that are discarded because the cache is full.

    • Concurrent waiters: Average number of threads that are waiting for a connection at the same time.

    • Number of managed connections: Specifies the number of ManagedConnection objects that are in use for a particular connection pool. This metric applies to Version 5.0 data sources only.

    • Percent maxed: Average percent of the time that all connections are in use.

    • JDBC time (milliseconds): Specifies the average time, in milliseconds, that is spent running JDBC calls. This time includes time that is spent in the JDBC driver, network, and database. This metric applies to Version 5.0 data sources only.

    • Average use time (milliseconds): Specifies the average time in milliseconds that a connection is used. The value is the difference between the time at which the connection is allocated and returned. This value includes the JDBC operation time.

    • Number of returns: Total number of connections that are returned to the pool.

    • Free pool size: Number of free connections in the pool.

    • Number of allocates: Number of connections that are allocated.

    • Average wait time (milliseconds): Specifies the average waiting time in milliseconds

  • System module: These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers. The system module has the following metrics:

    • CPU utilization: Since server start: Specifies the average CPU utilization since the server started.

    • CPU utilization: Last interval: Specifies the average CPU utilization since the last query.

    • Free memory (KB): Specifies a snapshot of free memory, in kilobytes.

  • Process module (for other servers): These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers. The process module has the following metrics:

    • Process resident memory (KB): Process resident memory, in kilobytes.

    • Process CPU utilization: Since server start: Process CPU utilization since the server start.

    • Process CPU utilization: Last interval: Process CPU utilization in the last interval.

    • Process total memory (KB): Process total memory, in kilobytes.

  • EJB module: These metrics can be used only on servers that are running WebSphere Application Server. The EJB module has the following metrics:

    • Average concurrent active methods: Specifies the average number methods that are active at the same time.

    • Total method calls: Number of calls to the remote methods of the bean.

    • Methods submodule: Method loads: Method load in the methods submodule.

    • Stores: Amount of time that the bean data was stored in persistent storage.

    • Message count: Number of messages that were delivered to the onMessage method of the bean. This message count applies to message-driven beans.

    • Average concurrent live beans: Specifies the average beans that are alive at the same time.

    • Removes: Number of times that beans were removed.

    • Returns to pool: Number of calls that are returning an object to the pool.

    • Passivates: Number of times that beans were moved to passivated state.

    • Gets from pool: Number of calls that are retrieving an object from the pool.

    • Drains from pool: Number of times that the daemon found the pool idle and attempted to clean the pool.

    • Ready count: Number of bean instances that are in ready state.

    • Average create time (milliseconds): Specifies the average time, in milliseconds, to run a bean create call. This time includes the time to load the bean.

    • Returns discarded Number of times that the returning object was discarded because the pool was full.

    • Activates: Number of times that beans were activated.

    • Server session usage (percentage): Percentage of the ServerSession pool that is being used. This metric applies to: message-driven beans.

    • Loads: Number of times that bean data was loaded from persistent storage.

    • Message backout count: Number of backed out messages that failed to be delivered to the onMessage method of the bean. This metric applies to: message-driven beans.

    • Methods submodule: Method response time (milliseconds): Method response time, in milliseconds.

    • Passivation count: Number of beans that are in a passivated state.

    • Pool size: Average number of objects that are in the pool.

    • Load time (milliseconds): Specifies the average time in milliseconds for loading the bean data from persistent storage.

    • Average remove time (milliseconds): Specifies the average time, in milliseconds, to run a beanRemove call. This time includes the time at the database.

    • Gets found: Number of times that a retrieve call found an available object in the pool.

    • Activation time: Average time, in milliseconds, for activating a bean object

    • Average drain size: Average number of objects that are discarded in each drain.

    • Methods submodule: Method calls: Specifies the number of method calls.

    • Destroys: Number of times that beans were destroyed.

    • Average server session wait time (milliseconds): Average time, in milliseconds, required to obtain a server session from the pool. This metric applies to message-driven beans.

    • Creates: Number of times that beans were created.

    • Average method response time (milliseconds): Average response time, in milliseconds, that elapses for remote method calls on the bean.

    • Instantiates: Number of times that beans were instantiated.

    • Store time (milliseconds): Specifies the average time, in milliseconds, for storing the bean data to persistent storage.

    • Passivation time (milliseconds): Specifies the average time, in milliseconds, for bean object passivation to occur.

  • Web application module: These metrics can be used only on servers that are running WebSphere Application Server. Web application modules have the following metrics:

    • Number of errors: Total number of times that an error was received from the servlet or JSP (JSP) files.

    • Total requests: Total number of requests that a servlet processed.

    • Response time (milliseconds): Specifies the average response time, in milliseconds, in which servlet requests finish.

    • Concurrent requests: Number of requests that are currently processing.

    • Number of reloads: Number of servlets that are reloaded.

    • Number of loaded servlets: Specifies the number of servlets that are loaded.

  • JVM runtime module: These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers. Java Virtual Machine (JVM) runtime modules have the following metrics:

    • Free memory (KB): Free memory, in kilobytes, in the JVM runtime.

    • Up time (seconds): Amount of time, in seconds, that the JVM has been running.

    • Total memory (KB): Total memory, in kilobytes, in the JVM runtime.

    • Used memory (KB): Amount of used memory, in kilobytes, in the JVM runtime.

  • Thread pool module: These metrics can be used only on servers that are running WebSphere Application Server. Thread pool modules have the following metrics:

    • Number of thread stops: Number of threads that are declared as stopped.

    • Percent maxed: Average percent of the time that all threads are in use.

    • Average active time (milliseconds: Specifies the average time in milliseconds that the threads are in active state.

    • Thread destroys: Total number of threads that are destroyed.

    • Pool size: Average number of threads in a pool.

    • Thread creates: Total number of threads created.

    • Concurrently hung threads: Specifies the number of threads that are concurrently stopped.

    • Number of cleared thread stops: Specifies the number of thread stops that are cleared.

    • Active threads: Number of threads that are concurrently active.

  
  Subexpression format for PMI Metric: From server start:

  PMIMetric_FromServerStart$moduleName$metricName operator LongValueL (with "L" suffix)

  
  Example:

  PMIMetric_FromServerStart$systemModule$cpuUtilization > 90L

  
  Subexpression format for PMI Metric: Last reported interval:

  PMIMetric_FromLastInterval$moduleName$metricName operator LongValueL (with "L" suffix)

  
  Example:

  PMIMetric_FromLastInterval$webAppModule$responseTime > 200L

• ODR server level metric: From server start

  We can use a subset of server level metrics that the on demand router (ODR) publishes. These metrics are cumulative and reported since the server start.

  Metric name

  We can use the following server level metrics:

  • Departs: Number of requests that are dispatched from the queue to the server during the reported interval. A request is considered to be dispatched just once, even if it fails at the first server and is retried at another time. The next event after dispatch is return. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Response time (milliseconds): Specifies an average response time for requests. To calculate this average, add the sum of the response times for requests that returned from the server to the client during the reported interval. The sum is in units of milliseconds. Divide by the serviced metric value to get the average response time. The response time of a request is the sum of the request's waiting time and service time. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Currently executing requests: Specifies the number of requests that are running at the end of the reported interval. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Service time (milliseconds) Specifies the average service time for requests. To calculate this average, add the sum of each request service time for requests that returned to the server during the reported interval. The sum is in units of milliseconds. Divide by the serviced metric value in the same interval to get the average. The service time of a request is the time from dispatch to return. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Wait time (milliseconds): Specifies the average wait time for requests. To calculate this average, add the sum of the time that each request spent waiting in the queue over the reported interval. The sum is in units of milliseconds. Divide by the number of departs to get the average wait time. Dropped requests do not contribute to this sum. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Errors: Number of requests that returned from the server with an error indicator during the reported interval. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Serviced: Specifies the number of requests that returned from server to client during the reported interval. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Timeouts: Number of requests that returned due to service timeout during the reported interval. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  
  Subexpression format:

  ODRServerMetric_FromServerStart$metricName operator LongValueL (with "L" suffix)

  
  Example:

  ODRServerMetric_FromServerStart$errors > 100L

• ODR server level metric: Last reported interval

  We can use the same set of metrics as the ODR server level metric: From server start operand. This operand uses an average of the reported values in the last interval. The interval is the length of the health controller cycle.

  • Departs: Number of requests that are dispatched from the queue to the server during the reported interval. A request is considered to be dispatched just once, even if it fails at the first server and is retried at another time. The next event after dispatch is return. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Response time (milliseconds): Specifies an average response time for requests. To calculate this average, add the sum of the response times for requests that returned from the server to the client during the reported interval. The sum is in units of milliseconds. Divide by the serviced metric value to get the average response time. The response time of a request is the sum of the request's waiting time and service time. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Currently executing requests: Specifies the number of requests that are running at the end of the reported interval. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Service time (milliseconds) Specifies the average service time for requests. To calculate this average, add the sum of each request service time for requests that returned to the server during the reported interval. The sum is in units of milliseconds. Divide by the serviced metric value in the same interval to get the average. The service time of a request is the time from dispatch to return. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Wait time (milliseconds): Specifies the average wait time for requests. To calculate this average, add the sum of the time that each request spent waiting in the queue over the reported interval. The sum is in units of milliseconds. Divide by the number of departs to get the average wait time. Dropped requests do not contribute to this sum. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Errors: Number of requests that returned from the server with an error indicator during the reported interval. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Serviced: Specifies the number of requests that returned from server to client during the reported interval. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Timeouts: Number of requests that returned due to service timeout during the reported interval. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  
  Subexpression format:

  ODRServerMetric_FromLastInterval$metricName operator LongValue (with "L" suffix)

  
  Example:

  ODRServerMetric_FromLastInterval$serviced > 10000L

• ODR cell level metric: From ODR start

  We can use a subset of cell level metrics that the ODR publishes. These metrics are cumulative and reported since the server start. We can use the following set of metrics:

  • Departs: Number of requests that are dispatched from the queue to the server during the reported interval. A request is considered to be dispatched just once, even if it fails at the first server and is retried at another time. The next event after dispatch is return. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Response time (milliseconds): Specifies an average response time for requests. To calculate this average, add the sum of the response times for requests that returned from the server to the client during the reported interval. The sum is in units of milliseconds. Divide by the serviced metric value to get the average response time. The response time of a request is the sum of the request's waiting time and service time. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Current queue length: Length of the queue at the end of the reported interval.

  • Service time (milliseconds)

  • Errors: Number of requests that returned from the server with an error indicator during the reported interval. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Average queue length: Specifies the average length of the queue. To calculate this average, add the sum of the queue lengths that are reported at each request arrival before insertion, and divide the sum by the number of arrivals. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Serviced: Specifies the number of requests that returned from server to client during the reported interval. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Timeouts: Number of requests that returned due to service timeout during the reported interval. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Currently executing requests: Specifies the number of requests that are running at the end of the reported interval. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Arrivals: Number of requests that arrived during the reported interval. The next event, if any, after arrival is either dispatch or drop. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Queue overflow drops: Specifies the number of requests that were initially accepted into the queue, and then ejected from the queue during the reported interval. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Queue drops: Number of requests that were initially accepted into the queue at some time and then were ejected from the queue at some time during the reported interval. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Delayed: Number of requests that arrived during the reported interval and were not immediately dispatched or dropped. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Wait time (milliseconds): Specifies the average wait time for requests. To calculate this average, add the sum of the time that each request spent waiting in the queue over the reported interval. The sum is in units of milliseconds. Divide by the number of departs to get the average wait time. Dropped requests do not contribute to this sum. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  
  Subexpression format:

  ODRCellMetric_FromServerStart$metricName operator LongValue (with "L" suffix)

  
  Example:

  ODRCellMetric_FromServerStart$arrivals > 10000L

• ODR cell level metric: Last reported interval

  We can use the same set of metrics as the ODR cell level metric: From server start operand. This operand uses an average of the reported values in the last interval. The interval is the length of the health controller cycle.

  • Departs: Number of requests that are dispatched from the queue to the server during the reported interval. A request is considered to be dispatched just once, even if it fails at the first server and is retried at another time. The next event after dispatch is return. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Response time (milliseconds): Specifies an average response time for requests. To calculate this average, add the sum of the response times for requests that returned from the server to the client during the reported interval. The sum is in units of milliseconds. Divide by the serviced metric value to get the average response time. The response time of a request is the sum of the request's waiting time and service time. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Current queue length: Length of the queue at the end of the reported interval.

  • Service time (milliseconds) Specifies the average service time for requests. To calculate this average, add the sum of each request service time for requests that returned to the server during the reported interval. The sum is in units of milliseconds. Divide by the serviced metric value in the same interval to get the average. The service time of a request is the time from dispatch to return. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Errors: Number of requests that returned from the server with an error indicator during the reported interval. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Average queue length: Specifies the average length of the queue. To calculate this average, add the sum of the queue lengths that are reported at each request arrival before insertion, and divide the sum by the number of arrivals. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Serviced: Specifies the number of requests that returned from server to client during the reported interval. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Timeouts: Number of requests that returned due to service timeout during the reported interval. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Currently executing requests: Specifies the number of requests that are running at the end of the reported interval. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Arrivals: Number of requests that arrived during the reported interval. The next event, if any, after arrival is either dispatch or drop. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Queue overflow drops: Specifies the number of requests that were initially accepted into the queue, and then ejected from the queue during the reported interval. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Queue drops: Number of requests that were initially accepted into the queue at some time and then were ejected from the queue at some time during the reported interval. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Delayed: Number of requests that arrived during the reported interval and were not immediately dispatched or dropped. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  • Wait time (milliseconds): Specifies the average wait time for requests. To calculate this average, add the sum of the time that each request spent waiting in the queue over the reported interval. The sum is in units of milliseconds. Divide by the number of departs to get the average wait time. Dropped requests do not contribute to this sum. These metrics can be used on servers that are running WebSphere Application Server or that are running other middleware servers.

  
  Subexpression format:

  ODRCellMetric_FromLastInterval$metricName operator LongValue (with "L" suffix)

  
  Example:

  ODRCellMetric_FromLastInterval$timeouts > 100L

• MBean operation metric: Long return type and MBean operation metric: String return type

  For Managed Bean (Mbean) operation metric operands, specify the Object name query string and the MBean method name.These metrics can be used only on servers that are running WebSphere Application Server.

  Object name query string

  When we create the Object name query string, escape all special characters with a backslash character.

  The value that you enter for the object name query string must have both the process=process_name>, and node=<node> strings specified, or none specified. If we specify both process=process_name>, and node=node>, the backend creates a singleton MBeanSensor sensor that senses the particular MBean on a server and node. If we specify none, the backend appends the name of the current server as the process name and the name of the current node as the node name, creating a MBeanSensor sensor for each server to which the health policy applies. If we specify only one of the twoprocess=<process_name> or node=node>, an error results.

  MBean method name

  Name of the MBean method to invoke.

  
  Subexpression format long metrics:

  MBeanOperationMetric_TypeLong$objectNameQueryString$methodName operator LongValueL (with "L" suffix)

  
  Example for long metrics:

  MBeanOperationMetric_TypeLong$WebSphere\:\*\,type\
    =HealthConditionLanguageInitializer\,node\=hipods3\,process\=nodeagent$getNumberOfOperands  > 10L

  
  Subexpression format for string metrics:

  MBeanOperationMetric_TypeString$objectNameQueryString$methodName operator StringValue

  
  Example for string metrics:

  MBeanOperationMetric_TypeString$WebSphere\:\*\,type\
    =HealthConditionLanguageInitializer\,node\=hipods3\,process\=nodeagent$getOperands  = 't'

• MBean attribute metric: Long return type and MBean attribute metric: String return type

  The MBean attribute metrics are used for querying an attribute of a MBean rather than invoking a method on the MBean. The operand takes the Object name query string and the attribute name as inputs. These metrics can be used only on servers that are running WebSphere Application Server.

  Object name query string

  When we create the Object name query string, escape all special characters with a backslash character.

  The value that you enter for the object name query string must have both the process=process_name>, and node=<node> strings specified, or none specified. If we specify both process=process_name>, and node=node>, the backend creates a singleton MBeanSensor sensor that senses the particular MBean on a server and node. If we specify none, the backend appends the name of the current server as the process name and the name of the current node as the node name, creating a MBeanSensor sensor for each server to which the health policy applies. If we specify only one of the twoprocess=<process_name> or node=node>, an error results.

  Attribute name

  Attribute that is queried on the MBean.

  
  Subexpression format for long metrics:

  MBeanAttributeMetric_TypeLong$objectNameQueryString$attributeName operator LongValue

  
  Example for long metrics:

  MBeanAttributeMetric_TypeLong$WebSphere\:\*\,type\
    =HealthConditionLanguageInitializer\,node\=hipods3\,process\=nodeagent$NumberOfOperands > 10L

  
  Subexpression format for string metrics:

  MBeanAttributeMetric_TypeString$objectNameQueryString$attributeName operator StringValue

  
  Example for string metrics:

  MBeanAttributeMetric_TypeString$WebSphere\:\*\,type\
    =HealthConditionLanguageInitializer\,node\=hipods3\,process\=nodeagent$OperatorList  =  'test'

• URL return code metric

  With this operand, we can ping any relative path (URI) on the server that is the target of this policy. The return value is used in the condition expression for the custom health policy.

  URL port number

  Port number to ping.

  URL relative path

  Specifies the URL to ping. Any special characters in the string must be escaped with a backslash (\) character.

  Value

  Specifies an integer that is the expected return code of the ping.

  Use this operand to ping any general purpose URL by selecting the on demand router (ODR) as the target of the health policy and by setting the appropriate routing rules in the ODR.

  We can use this operand to select members that are running WebSphere Application Server, or that are running other middleware servers.

  
  Subexpression format:

  URLReturnCodeMetric$portNumber$relativePath operator IntValue

  
  Example:

  URLReturnCodeMetric$9060$ibm\/console\/login\.do = 200

  The URL sensor returns0 if the Web site cannot be reached:

  URLReturnCodeMetric$9060$ibm\/console\/login\.do = 0

• External URL return code metric

  With this operand, we can enter an absolute URL instead of a relative URL. By doing so, we can periodically send a ping request to other targets than application servers or on demand routers.

  Always create a custom health action without specifying a target server.

  
  Example:

  ExternalURLReturnCodeMetric$http://foo.bar.com <> 200

  If http://foo.bar.com returns a different response code than 200, the health policy is triggered. Take this custom action to run a custom script, for example a script to recycle a web server.

Operator

• Equals (=): The equality operator expresses a case-sensitive match.

• Not Equals (<>): The not equal operator expresses that the operand value is not equal to the value you enter.

• Greater Than (>): The greater-than operator is for use with numbers.

• Greater Than or Equals (>=): The greater-than or equal to operator is for use with numbers.

• Less Than (<) : The less-than operator is for use with numbers.

• Less Than or Equals (<=): The less-than or equal to operator is for use with numbers.

• Between (BETWEEN): The value must be between a Lower bound and Upper bound specified.

• In (IN): The value must be in a list of values. We can type in values and add them to a list.

Value

Depending on the operator that you choose, type in a value for the subexpression to create.

Subexpression

After you click Generate subexpression, this field displays the generated subexpression fragment based on the options selecteded. To add this subexpression to the custom health condition, click Append.