As CirrusSearch maintainer I want the cirrusSearchCheckerJob jobs to be handled in time so that they don't pile-up indefinitely.
Since the switch to eqiad the cirrusSearchCheckerJob seems to be slower, the rate of ids being checked has dropped drastically just after the switch:
The impact is that the cirrusSearchCheckerJob jobs do pile up.
AC:
- find the cause of the slowdown
- the backlog of cirrusSearchCheckerJob jobs should remain sane
- re-enabled the saneitizer after resolving performance issue(s)