A monitor model produces several CWMRT2007W warning messages in the SystemOut.log file in the server or cluster where the moderator EJB for the model is running; eventually the model may completely stop processing inbound events.
Warning messages similar to the following appear in the SystemOut.log file on the server (stand-alone environment) or the cluster members where the moderator EJB for the model is deployed (network deployment environment):
[12/10/10 2:43:22:410 EST] 00000058 ConsumerDaemo W com.ibm.wbimonitor.om.ClipsAndTacks.20100722090000.runtime.moderator.ConsumerDaemonSerialST issueTimeBasedEvent() CWMRT2007W: It took more than a third of the time based triggers check interval to evaluate the time based triggers. This might cause a performance degradation
Monitoring Model ID: ClipsAndTacks
Monitoring Model Version: 20100722090000
Execution Time (so far): 181
++++ Check runtime properties to increase the time based triggers check interval.
Evaluating time-based triggers cuts into the time available for regular inbound event processing. If this evaluation happens frequently, and the incoming event rate does not stay relatively low, a performance degradation can result. It is also important to note that the more active instances there are for the monitoring contexts in which time based triggers are defined, the longer each time based trigger evaluation cycle (frequency defined by the Recurring wait time check interval property in the model runtime configuration) will take to complete.
Resolving the problem
The Recurring wait time check interval can be increased in the model runtime configuration via the WebSphere administrative console. Refer to this Information Center link (for Version 7.0, but this information also applies equally to Versions 6.1.2 and 6.2):
Note that step 9 of the procedure states that the monitor model application must be restarted for the property change to take effect.
This property should not be set to a value higher than the smallest evaluation interval for all time based triggers defined in the monitor model, to ensure that all triggers are evaluated frequently enough.