Log Event Metrics on Linux are causing high CPU usage
Last Updated June 20, 2013
Log Event Metrics applied to Linux based computers can cause moderately high CPU usage (approximately 30%). This occurs when multiple Log Event Metrics are applied to Linux machines. The issue is generally not seen unless the monitored log is being written to at very high rates and simultaneously being read by the Log Event Metric.
There is no associated error message. The only indication of an issue is that CPU usage may increase when Log Event Metrics are applied.
The root cause is unknown at this time. When Log Event Metrics are applied to a particular Linux based computer that is continuously writing to the log being monitored while the Log Event Metric is reading from it, CPU usage will increase slightly (4-6%). It looks like this issue may be compounded when other metrics are applied, but more research is required to confirm that fact.
There is no solution or workaround for this issue at this time, but it is recommended to reduce the rate of logging to the particular log file being monitored if possible. Logs that are written to at lower rates tend to not experience the high CPU usage issue.
All supported Linux Distributions: Red Hat Linux 3 - 6.1, Novell SUSE Linux 8 - 12, Sun Solaris 7 - 11.
Etrack information for this issue.
Imported Document ID: TECH207268
Subscribing will provide email updates when this Article is updated. Login is required to Subscribe