probe crashing repeatedly does not trigger max restarts limiter

Products

• DX Unified Infrastructure Management (Nimsoft / UIM)

Issue/Introduction

Ordinarily after 10 restart failures, a probe goes into a max_restarts condition and turns red, but in some cases, you have observed that a probe restarts continuously and never triggers the restart limiter.

Environment

DX UIM - Any Version
any probe - any version

Cause

Slow startup of probe / brief runtime before restart

By default, the max restarts functionality does not kick in unless the subsequent probe restarts occur within 15 seconds of one another.

If a probe starts and runs for more than 15 seconds before restarting, or the probe startup takes more than 15 seconds before the probe fails and restarts again, the restart will not be counted toward the max_restarts limit.

Resolution

To adjust this, a configuration key may be added to the robot.cfg in the main <controller> section:

default_fail_window = ##

where ## is a value in seconds.

For example, if you want any restart within 60 seconds to be counted toward the limit instead of 15 seconds, add the key/value:

default_fail_window = 60