The SlowOperationDetector monitors the operation threads and collects information about all slow operations. An Operation is a task executed by a generic or partition thread (see Operation Threading). An operation is considered as slow when it takes more computation time than the configured threshold.

The SlowOperationDetector stores the fully qualified classname of the operation and its stacktrace as well as operation details, start time and duration of each slow invocation. All collected data is available in the Management Center.

The SlowOperationDetector is configured via the following system properties.

  • hazelcast.slow.operation.detector.enabled
  • hazelcast.slow.operation.detector.log.purge.interval.seconds
  • hazelcast.slow.operation.detector.log.retention.seconds
  • hazelcast.slow.operation.detector.stacktrace.logging.enabled
  • hazelcast.slow.operation.detector.threshold.millis

Please refer to the System Properties section for explanations of these properties.

Logging of Slow Operations

The detected slow operations are logged as warnings in the Hazelcast log files:

WARN 2015-05-07 11:05:30,890 SlowOperationDetector: []:5701
  Slow operation detected:
  Hint: You can enable the logging of stacktraces with the following config
  property: hazelcast.slow.operation.detector.stacktrace.logging.enabled
WARN 2015-05-07 11:05:30,891 SlowOperationDetector: []:5701
  Slow operation detected:
  (2 invocations)
WARN 2015-05-07 11:05:30,892 SlowOperationDetector: []:5701
  Slow operation detected:
  (3 invocations)

Stacktraces are always reported to the Management Center, but by default they are not printed to keep the log size small. If logging of stacktraces is enabled, the full stacktrace is printed every 100 invocations. All other invocations print a shortened version.

Purging of Slow Operation Logs

Since a Hazelcast cluster can run for a very long time, Hazelcast purges the slow operation logs periodically to prevent an OOME. You can configure the purge interval and the retention time for each invocation.

The purging removes each invocation whose retention time is exceeded. When all invocations are purged from a slow operation log, the log is deleted.