Enterprise Administration Guide › Viewing Your Enterprise Implementation › View Endpoint Status › Monitor Endpoint Memory Usage and Responsiveness

Monitor Endpoint Memory Usage and Responsiveness

An endpoint restarts itself when the process exceeds critical memory thresholds, or (on Windows only) if the process uses too many handles. The application dumps the process memory and information on stuck processes for further analysis.

Edit the seoswd section of the seos.ini configuration file to configure timeouts, schedules, and thresholds of the watchdog.

Endpoints can also detect slow responsiveness or whether the authorization engine process hangs (seosd).

• Slow Windows endpoints give full bypass to system events and retry recovering periodically.
• Slow Unix endpoints restart seosd, give bypass for system events coming from a suspected process, and dump a log to the file system.

Note: In bypass mode, policy handling is temporarily reduced.

Edit the seosd section of the seos.ini configuration file to turn off the automatic bypass for programs that are suspected to cause deadlocks. Deadlock detection is automatic and non-configurable.

You can monitor automatic process restarts and memory usage as follows:

Follow these steps:

1. Consult the following logs:
   • System log (syslog for Unix and Event Viewer for Windows)
   • Product log (ControlMinder audit file)
   • DMS event
   • Endpoint status in World View
2. Monitor endpoint memory usage through the secons utility:

   secons -i 
   

   The utility displays counters for virtual memory, physical memory, and process handles.

More information:

seosd