|Thierry Carrez||Jun 21, 2012 7:29 am|
|Mark McLoughlin||Jun 21, 2012 7:58 am|
|Thierry Carrez||Jun 21, 2012 8:31 am|
|Vincent Untz||Jun 22, 2012 1:55 am|
|Martin Packman||Jun 22, 2012 7:37 am|
|Vishvananda Ishaya||Jun 26, 2012 2:44 pm|
|Thierry Carrez||Jun 28, 2012 2:32 am|
|Alan Pevec||Jul 2, 2012 4:41 am|
|Thierry Carrez||Jul 2, 2012 5:02 am|
|Subject:||[Openstack] [Nova] How to improve our bug triaging ?|
|From:||Thierry Carrez (thie...@openstack.org)|
|Date:||Jun 21, 2012 7:29:39 am|
This is a follow-up on a recent discussion at the weekly project/release meeting. At the beginning of the month we had a BugTriage day that allowed us to make the Nova bug database much more relevant and triage a lot of incoming new bugs. However, since then, the numbers went up again:
This is the number of New/Undefined-priority bugs, which typically are bugs that were submitted by users that we don't even look at. There are more bugs filed (for example by developers) but they usually get triaged and solved alright. The leftovers number grows by about 50/month if we don't do anything about it.
Having people report bugs for our software is a chance, and we should not discourage anyone by not giving the resulting reports the minimal attention they deserve. Furthermore, regular gardening of the database ensures that it stays current and we can use it to prioritize fixes correctly. However it cannot be a single-person job, given the volume.
So how do we, Nova developers, collectively solve this over the long run ? Suggestions include:
* Run BugTriage days more often ? We could have regular (monthly?) Nova Bugtriage days to get rid of what accumulated in the mean time. But I fear that urgent bugs might not get the attention they deserve, and that over time less and less people participate to those exciting events...
* Add BugTriaging to the core dev duties (new ReviewDays task) ? May be a good idea since core devs are also Nova bug supervisors, so they actually /can/ triage bugs. However it sounds a bit unfair to core devs, and I fear it would result in Nova reviews getting even less attention than they get currently...
* Should we just encourage more people to do BugTriaging ? Sounds like the obvious solution. However to do good triaging, you need bug supervisors rights, and when I proposed to open that team (the nova-bugs team) to anyone, there was resistance. Maybe it makes sense for Nova though... or maybe someone should actively manage that team and grant rights to any known triager that needs them.
Thoughts ? Other suggestions ?
-- Thierry Carrez (ttx) Release Manager, OpenStack