On May 24, 2007, at 10:25 AM, Basil Shubin wrote:
I've setup PerfConf to use check_http for site checking, commands
looks like this:
check_http -H google.com -t 120 -u /index.html -s "Google"
But very often I got notifications about that command timed out.
So, my questions are:
1. How I can fix this, i.e. command shouldn't 'timed out'? What I
need to do this, increase value for -t option?
2. Can I just ignore them, i.e. configure event notifications to
ignore messages like this one? What possible drawbacks of this
approach?
3. Any other idea and suggestions?
BTW, what are possible causes of this timed outs?
Basil,
Those timeout notifications are being caused by the site not
returning data in less than 2 minutes. You could increase the -t, but
you'd probably want to know if the site took that long to respond.
You can ignore the events if you want. Take a look at the event
mapping section of the admin guide for details on how to do this.
Normally you'd just want to match the event as specifically as
possible and map it to the /Ignore event class to have it dropped.
There are so many possible causes for timeouts. DNS resolution
failures (try specifying an IP instead of hostname to the -H option
to eliminate this possibility), firewall problems, or any problem
along the path from your Zenoss server to the destination web server.
My suggestion would be to not ignore these events, but troubleshoot
the issue that is causing them.
Chet Luther
[EMAIL PROTECTED]
_______________________________________________
zenoss-users mailing list
[email protected]
http://lists.zenoss.org/mailman/listinfo/zenoss-users
