Page 1 of 1

Disk IO report on load alert

Posted: 29 Aug 2011, 20:44
by amoon
I was wondering if the developers of this fine script would consider the option of including iotop output with the other load alert reports.

A lot of times when we get load alerts it's because of disk IO and not process CPU usage, so this would really help us track things down.

I know there is a "LOAD_ACTION" value which allows you to execute a custom script on the high load event, but whenever it tries to execute I just end up getting a bunch of python errors in my messages.

My loadaction script:

#!/usr/bin/env bash
iotop -b -n1 2>&1 | mail -s "$HOSTNAME iotop output" incidents@domain.tld

//end

Re: Disk IO report on load alert

Posted: 30 Sep 2011, 09:17
by chirpy
You really need to resolve the errors that you are seeing in your script as that is what LOAD_ACTION is for. You will need to ensure that you set the environment up correctly in your shell scripts (e.g. the PATH variable if necessary) to ensure it runs as expected.