Nagios in Vicidial

[baka-na]

Hi Vicidialers,

Would there be any conflict if I install nagios plugins and nrpe on a Vicidial Server? Have anyone tried this?

Thanks a lot in advance!

[williamconley]

Only in that Nagios uses system resources. I've not actually seen a server with nagios LIVE, but I've seen several with it installed.

Nagios is kind of overkill for Vicidial though. All you really need to watch is that mysql is running and the db is not corrupt and asterisk is running and accepting calls.

Nagios may verify running mysql and asterisk ... but I don't think it'll verify that there are no corrupt tables or that asterisk can make/take calls.

[baka-na]

hmm..ok

Thanks a lot for the input William. I also worried if the installer/package of Nagios will have conflict to those of Vicidial/asterisk/etc...

Best Regards.

[williamconley]

I've seen several servers with nagios installed, vicidial still seemed to be quite functional and unencumbered.

But don't fool yourself that a "pass" from nagios means anything if it does not verify asterisk is actually taking calls and mysql tables are not corrupted. False sense of security and all ...

[boybawang]

try opmanager

[williamconley]

Details on the Free edition seem ... sparse. LOL

And I do not see anything that indicates it will ensure mysql and asterisk are actually working (ie: no corrupt tables for mysql and Taking Calls for asterisk ...).

Just being sure processes are running can be handled with any one of several apps (easiest of which being Webmin). But if a mysql table corrupts or asterisk freezes ... the warnings won't go out and a client will be seriously upset. We monitor both of those to allow us to fix a server before a client can complain. Merely checking to see the processes have a PID doesn't mean they are working as designed.

We stop short of actually generating a phone call for most clients, but we have a couple who insist on having calls generated regularly to ensure everything works. (So our system calls their system, their system answers and logs the call, and then our system checks the log to be sure all went as expected ...). That covers numbers being ported unexpectedly, asterisk freezing, carrier routing failing ... and of course we check for crashed tables regularly for those on the deep watch.

All custom of course, no "app" with a two-week learning curve. LOL

[Acidshock]

I use Icinga myself. Basically a nagios fork/rebuild. I have several scripts that check space, check to make sure registration is possible, etc. It is a little hard though to confirm its placing calls 100%.

It is possible to do though if you can place a dialplan that creates a file. You then tie a did to that extension and have your script dialout to the did. Then when the call is received it creates the file and hangs up. You then have your script dial the extension then check to see if a file exists, if it does it deletes the file and passes the check.

[williamconley]

Easier to create a DB entry (like ... a log?) and then check to see if the log entry was created. In fact, any existing DID will already do that. You just need to create the calls and check to see if they "happened". On a high-volume system, this won't add enough to the cost to be noticed ... but in case of a problem it is always cool to find/fix before the client even knows there is a problem. (Nothing beats having the manager call you just AFTER you fix the problem and he stops in mid-sentence to say "nevermind, it's working now ... thanks!")