Filing this bug after a short conversation with gmb:
<Ursinha> gmb, tell me, of all of those errors in the checkwatches' summary, which ones we can do nothing about?
<gmb> Ursinha, Most of them, in fact. What we'd ideally do about them is suspend the bug watches because they don't work, and for the ones that error a lot that's pretty much what we'll do; they'll be checked at most once every 7 days.
<gmb> Ursinha, I'll take a look see if we've got any bugtrackers that are consistently badly behaved; we should probably stop updates on those.
<Ursinha> we want to get rid of false errors so we can rely more on our reports
<gmb> Ursinha, Agreed. It's a little hard though, because a lot of the errors are genuine but we can't do a lot about them because the problem is on someone else's system.
<Ursinha> gmb, in these cases, can we handle them gracefully in lp's side, or make them only informational?
<gmb> Ursinha, We already have a backoff mechanism in place. However, now that we're tracking BugWatchActivity we can probably stop recording OOPSes for some things.
<gmb> Ursinha, Ideally, we need to do more work on checkwatches but we've got to finish the subscriptions work before we can consider it.
<Ursinha> gmb, for now, do you think disabling recording some oopses is possible? by possible I mean it won't just cover the mess, but they can be disabled safely
<gmb> Ursinha, I'll take a look and let you know. Can you file a bug to the effect of "Checkwatches produces OOPSes for non-urgent items" or some such? I'll pick it up tomorrow.
Bug 390543 is related.
I think this is less of a problem now that the bugs OOPS report doesn't include checkwatches OOPS. So I'm marking this low and removing the OOPS tag, since it will fall under general improvements to how we handle checkwatches errors. If I'm misunderstanding the importance of this, please let me know.