585737 - Add nagios monitors to middleware on dm-bp-mware01

Status: RESOLVED FIXED

(mozilla.org Graveyard :: Server Operations, task)

Description

[Laura Thomson :laura]

This powers the webapp and at present has zero monitors.

Please check:
- Apache is running
- that you are getting 200s from the services (URL to come)

Comment 1

[Ryan Snyder [:ryansnyder] [:rsnyder] [:rysny]]

Here's a list of the current API calls that we provide:
http://code.google.com/p/socorro/wiki/APICalls

{url_prefix} for any of these calls will need to replaced with the url for the webservice layer, which can be found in webapp-php/application/config/webserviceclient.php -> $config['socorro_hostname']

Both the ADU and Topcrasher calls are currently available.  The Crash Report calls will be available in 1.8.

The easiest call to prepare the URL for would probably be the ADU call.  

http://example.com/201005/adu/byday/p/Firefox/v/3.6.7/rt/crash/os/Windows/start/2010-08-01/end/2010-08-08

You can replace the current version "3.6.7" with the current major version, which can be derived from the productdims table.  Replace the end date with today's date and the start date with this day last week.

Comment 2

[K Lars Lohn [:lars] [:klohn]]

rather than have a call that would have to be maintained by keeping version and time windows up to date, it might be better to just have a call that says, "yes, I'm healthy".  If that service works, we can infer that the other services are working.  

I actually thought there was a 'hello' service on there that could have served this purpose, but it appears to have disappeared over the months...

Comment 3

[Ryan Snyder [:ryansnyder] [:rsnyder] [:rysny]]

A "yes, I'm healthy" call sounds much better!

Comment 4

[Aravind Gottipati [:aravind]]

I'd added generic host health checks to the box, but I'd like a better call to check for the service health.  The 'replace the current version "3.6.7" with the current major version' doesn't work well with http string checks in nagios.  We'd have to invent a custom check and one that would keep changes as the app changes.

I'd a simpler check please.

Comment 5

[Aravind Gottipati [:aravind]]

We now have a check for the url in comment 1.