Randomly receive "UNKNOWN - No data found in agent output"

Hi,
I receive randomly “UNKNOWN - No data found in agent output” from our local checks. It can’t be a runtime problem, because these checks shouldn’t run for long. They just check if a process is still running.

How can I prevent that or ignore it?

Kind regards
thuranga

I have encountered this issue when a remote node goes to high load the nrpe or other remote agent can not execute the check properly and then times out and returns the error you are getting.
if you can correlate the errors to high activity on the server - this might be the issue.

Are you using the check_procs plugin for this or some custom plugin?

1 Like

No, there is no high activity on these servers …

And I am not using the check_proc plugin. I use a custom shell script for evaluating the informations, because I need more than a “process running”. E.g.: I check if gitlab-runner is running and print out the version and number of concurrent runner

Yes, it is true in my case. Where timeout value is defined? Maybe I could increase this value?

By default the value is 300 seconds and is defined in the plugin object definition.
you can override it with -t on the service command definition

The other side is the nrpe daemon on the remote node itself, and the default timeout on it is usually 30 seconds.
To change that you need to edit the nrpe.conf file.

/etc/nagios/nrpe.conf

Thanks but I’m not using nrpe. I use standard check_mk_agent and I guess that more than 5 sec. is too much.
It is strange that Unknown only occurs on services with WARN status.