Number of cached events is too high, the oldest have been clamped

Document ID : KB000092724
Last Modified Date : 23/04/2018
Show Technical Document Details
Issue:
We have received constant entries of following warning message in the EM log every second, reaching >62k entries in less than a day:

[WARN] [pool-30-thread-1] [Manager.AppMap] Number of cached events is too high, the oldest have been clamped!

What does this warning message imply and what could have been triggering it?


 
Environment:
CA Application Performance Management (CA APM) 10.x
Cause:
As indicated by the text, this warning means there are too many events in the cache, waiting to be processed by the Team Center.

The default clamp value is 100,000. It can be explicitly configured by adding the introscope.apmserver.events.clamp property into the IntroscopeEnterpriseManager.properties with the desired value. The cached events are AppMap related events, which could consist of Transaction Tracing, Alerts or others.

These events could pile up in the cache if the Team Center or the APM Database is performing poorly.

In this specific case, it was a result of the Enterprise Team Center (ETC) / Master EM not starting successfully due to the reason discussed in this KB article.
As the ETC was not up and running, cached events in providers were piling up due to events not getting processed & cleared off the cache.
Resolution:
Fixing the Enterprise Team Center startup issue has restored the events processing, which in return, cleared the events off the cache & ceased the warning message.
Additional Information:
Enterprise Manager failed to start due to "Failed to map segment from shared object"
Enterprise Team Center