Alert – Process dead

[StanSmith PA]

Hi All

I’m developing and testing new alerts and can’t get one to trigger – so I’m not sure if I’ve built it correctly.  I’m attempting to send an alert email when a process has been dead for 5 minutes.  This looks good in the Alert Configurator, but never fires in testing.  To test – I manually stop a process with hcienginestop and wait.

Suggestions?  v5.8, Linux Red Hat

Running two environments summit_health & summit2 – double-checked that this default.alrt is running in summit_health

Thanks,

Stan

Code:

{ALERT
   { NAME 11-ProcessDown }
   { VALUE pid }
   { SOURCE {
       {
           { PROCESS Medi_PhaCtl }
       } {
           { PROCESS Med_Lab }
       } {
           { PROCESS adt }
       } {
           { PROCESS medi_omni1 }
       } {
           { PROCESS cpal }
       }
   } }
   { WITH -2 }
   { COMP {== dead} }
   { FOR {nmin 5} }
   { WINDOW */*/*/* }
   { ACTION {
       { exec {echo “CRITICAL: a summit_health process is down for 5 minutes.” | mail -s “%A” abunchaemails@summithealth.org} }
   } }
}

Topic

[Levy Lazarre]

Stan,

I am on a lower version of Cloverleaf, so things may have changed. If I were trying to set up this type of alert on my system, I would need more asterisks (*) in the Time window to indicate that the offending event could happen any time.

*  Seconds

*  Minutes

*  Hours

*  Days

*  Months

*  Weekdays

You only have 4 asterisks, so it doesn’t seem that you have a setting for the Months, to signify that it can happen any month.

Maybe you should try that and see what happens.

[StanSmith PA]

Thanks Levy

It was the Time Window and…

As it turns out 5.8R1 has an issue with

{ COMP {== dead} }

which does not work.

{ COMP {!= running} }

does work, and gets the job done.  (Fixed in R4)

Thanks,

Stan

[Author]

Replies