Time to merge over time

[Hixie]

trafficstars

It would be really interesting to see metrics like the mean time to merge (and maybe the standard deviation?) over time for a repository, organization, or person.

[TristanCacqueray]

According to https://stackoverflow.com/a/51726136 , it might be possible to update the current query: https://github.com/change-metrics/monocle/blob/90191a2fbe0fdc4c89982bdfa0c2c69ce313787a/monocle/db/queries.py#L556-L564

[morucci]

@TristanCacqueray do you think we should consider a more general trends system where a trend is configured then computed on demand ? It seems this sort of thing will be intensive for the DB if computed/displayed by default.

[TristanCacqueray]

It might be difficult to cache the result for every filters, we might have to restrict them to project definition. But let's give avg_bucket a try first, perhaps it is not that intensive.

More generally, we might want to integrate memcache to improve response speed, e.g. for sprint review.