Page MenuHomePhabricator

Configure Nagios to report to toolserver IRC for toolserver outages
Closed, DeclinedPublic

Description

Author: fearow00

Description:
It would be incredibly useful to add a second contact to nagios to report to #wikimedia-toolserver on freenode when any of the services in the toolserver groups status changes.

The groups to monitor:

  • toolserver_database
  • ts_ext_store

It would really help us to see that there was a server problem etc and know that it's not an issue in our scripts etc.

I wasn't sure where to ask, but since it is regarding nagios which is part of the main Wikimedia setup rather than the toolserver setup, this seemed like the most appropriate place to request.

Thanks,
Matt


Version: unspecified
Severity: enhancement
URL: http://nagios.wikimedia.org

Details

Reference
bz10998

Event Timeline

bzimport raised the priority of this task from to Lowest.Nov 21 2014, 9:51 PM
bzimport set Reference to bz10998.
bzimport added a subscriber: Unknown Object (MLST).

fearow00 wrote:

Any comments, any idea when/if this will be implemented?

bugs wrote:

Ask River (flyingparchment). Contact the ts-admins {at] wikimedia.org.

fearow00 wrote:

I was under the impression that nagios was a part of the main cluster, and the toolserver staff wouldn't be able to do it without asking a main dev anyway.

bugs wrote:

(In reply to comment #3)

I was under the impression that nagios was a part of the main cluster, and the
toolserver staff wouldn't be able to do it without asking a main dev anyway.

River is both a toolserver and Wikimedia server root admin. Therefore he would be the best to contact about this, being able to configure it properly on both ends and explain if or if not it is needed.

fearow00 wrote:

I have talked to River, and they said to talk to Jeluf. Also, it's only the nagios end that needs configuring.

jeluf wrote:

nagios is not talking to irc. It's some strange script that does this. I don't see any well-maintainable way to tell nagios-wm which events it should report where.

I'd recommend to install nagios on one of the toolservers.

I've connected nagios-wm-echo bot to #wikimedia-tech and #wikimedia-toolserver which echoes all toolserver-related nagios-wm messages from #wikimedia-tech to #wikimedia-toolserver.