old bugs in performance backlog

[Ben Kelly]

Hi all,

I'm trying out my first rotation since triaging became automatic.  While looking at the second step about regression backlog it occurred to me there are a lot of very old bugs there.  Are any of those actionable?  Should we just bulk close any performance bugs that are more than 6 months old?  It seems at some point they just are accepted as part of our new performance baseline.

Closing bugs that have not been modified in 180 days would cut the backlog in about half; from ~1600 to ~800.  Closing bugs that have not been modified in 90 days would bring the backlog down to ~500.

If we do keep these open, do folks have any guidance on how to make progress on them?

Thanks.

Ben

[Timothy Dresser]

+Performance Regression Sheriff Dogfood 

An additional signal we may want to use is regression magnitude.

There's a dashboard of large system health bugs here.

Maybe closing all bugs > 90 days old that aren't > 5%, or something like that, would be a reasonable compromise?

[Ben Kelly]

For old bugs with assignees we could also ping them to see if they think it is still actionable or not.

[Timothy Dresser]

A one time ping followed by closing if there's no activity makes sense to me. 

+Matthew De Luca - how hard is that in practice?