We need a public place to announce downtimes, maintenance, and changes to the system. Can we set something up on blog.mozilla.com or similar?
There is http://status.mozilla.com/ which carries the mozservices twitter feed, perhaps Socorro can hook into there too.
what URL would you like to check for socorro?
Note that adding it to status.mozilla.com is fine and will cover direct downtime, but will not cover other issues like maintenance and changes to the system. Often times, those latter two issues cause problems but not direct downtime.
Let's do both. Jeff, what do we hook it up to?
(In reply to comment #4) > Jeff, what do we hook it up to? I'm not sure what you're asking. Basic checking is just a URL (it's smart -- wants a 200, etc). I can also add regex string validation, etc, if needed. It's hosted by Watchmouse, and thus has a fairly robust check system, so let me know if you want to do anything fancy.
Question for jv: Can we do multiple services? Socorro has quite a few moving parts, and it might be interesting to have different things, like crash-reporting - working. crash-stats frontend - working, hbase - not working, etc. Question for laura: to get a valid status of the overall system, perhaps having a few checks in place, one url check that requires the system to pull a crash from hbase, one that hits the crash-reports /submit url with a post and one that just looks for a 200 on the main page. I think we have enough simple ways to do checks that this shouldn't be too hard to implement. If we can just have on url check, then I'd suggest just having it try to load a url that requires a crash to get pulled from hbase, so that the "overall status" would either be good or bad depending on if the "overall system" is working.
(In reply to comment #6) > Question for jv: Can we do multiple services? Socorro has quite a few moving > parts, and it might be interesting to have different things, like > crash-reporting - working. crash-stats frontend - working, hbase - not working, > etc. as many as you want. Keep it down to a few dozen, please, else things get ugly :) we also have a secret "Experimental" status page if you need to try stuff / work out thresholds. -- http://status.watchmouse.com/11695/
I'll grab this bug. Since I sit across a cube wall from jv, I think I can get the support I need in setting this up.
I've created http://blog.mozilla.com/socorro with laura and rhelmer as administrators. I'll work on getting watchmouse set up on status.mozilla.com next week.
The blog part is done. I created bug 638421 for tracking the status.mozilla.com setup.