The process check doesn't allow to count matching processes, in this case you can use the "check program" with simple script, that will pgrep chrome and return the processes count as exit value, so you can then use "if status < 7 then alert"
Thanks for the input.
This indeed is listing a total of 7 matches. Any idea how to force a action of any of the 7 have a problem?
Any thoughts about option 1? How to fix the ‘recurring action’?
Op 4 nov. 2015, om 22:01 heeft Martin Pala <address@hidden
> het volgende geschreven:
Using match based "check process" should be possible, but you need unique pattern - if multiple processes match the pattern, monit will report that the process is running.
You can test the pattern using:
monit procmatch "chrome"
On 04 Nov 2015, at 10:19, Joost Plas | Pandora Producties.nl <address@hidden> wrote:
I’m looking for the best practice to monitor the health of Google Chrome.
Am I right to understand that Chrome doesn’t have a PID file?
I’ve tried the following 2 techniques:
check program chrome with path /usr/bin/google-chrome
if status!=0 then exec “/usr/bin/pkill chrome”
(I have a script that auto reboots chrome with the correct settings when I kill chrome).
This is sort of working when I simulate a crash (type in chrome://crash in browser). The problem is that it keeps killing chrome every time monit runs.
I think it doesn’t reset the last exit value? When I type ‘monit status’ it keeps saying ‘last exit value = 1’. How do I fix this?
Also I keep getting errors in M/Monit ‘cannot open display’. I think this is because monit keeps trying to reopen chrome by itself. I thought I could fix this by adding ‘mode passive’ but this doesn’t seem to be allowed here because I get a syntax error.
check process chrome matching “chrome”
The problem with this version is that nothing happens when I manually crash chrome. It keeps saying ‘status = Running’.
Hope someone can help me. I’m basically looking for the best technique to catch all possible issues with chrome. Thanks!