>> And I want to say folks, I feel your pain. I come from a QA and >> operations background in large, mission critical, distributed systems >> and logs have always been my best friends **and** the bane of my >> existence when they don’t have what you need. I get it. I want to help >> fix it.
> I'm also looking forward to cleaning up the logging in OpenStack projects. > May I suggest on the etherpad that we take an approach of listing > *specific* log messages that we (devs doing debugging and operators > doing diagnostics/operations) find less than useful? > I think if we keep the etherpad focused on specific log messages, we can > then start to identify: > * changes to those log messages (structure, level, audience, payload, etc) > * log message "archetypes" that we can then use to generalize into > best practice documentation on the wiki (to add to what is already there > [1]) > Does this sound like a reasonable approach? +1 We recently setup logstash here and doing the grok magic was quite a pain with all the different way's the logging is formatted. We also throw a bit in the bin because there is no useful info in it. Looking at our logstash grok I could probably make some suggestions on what we find useful and not :) Cheers, Robert van Leeuwen _______________________________________________ OpenStack-operators mailing list [email protected] http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-operators
