Hi,
On Wed, Nov 19, 2014 at 04:58:37PM +0100, Christian Aistleitner wrote:
On Wed, Nov 19, 2014 at 12:52:05PM +0100, Gilles Dubuc wrote:
I saw mentions of EventLogging DB problems in earlier threads, could it be that a lot of events didn't get recorded or something similar?
[...]
I'll have a look.
It turned out that on 2014-11-14 event volume grew beyond what EventLogging's database writer could take.
https://wikitech.wikimedia.org/wiki/Incident_documentation/20141114-EventLog...
An improvement of the database writer got deployed 2014-11-20T21:00, and it could handle the volume since then easily. But the fix still has to pass today's busy hours.
This was a limitation of EventLogging's database writer, and not EventLogging itself. So the secondary logs that EventLogging keeps should be good. I'll see to get the database backfilled from them and update the above Incident documentation page accordingly.
Have fun, Christian