Gitlab Collector: Stuck idle in transaction collector and test

[Sticksman]

Split from #819

[tomhughes]

Is hard coding a specific time limit in the collector really a good idea?

We've been using a modified version of the pg_process_idle query which includes both idle and idle in transaction and breaks the metrics down by both state and application which then allowed us to write alerts on pg_process_idle_seconds which look for excessive delays.

[SuperQ]

@tomhughes Yea, you're correct, it would be good to have a flag for the hard-coded values.

Perhaps you would like to contribute your version as a new collector method?

[SuperQ]

@tomhughes Another thing I was thinking about was if we wrote a pg_stat_activity collector that provided several different metrics. Possibly dump a raw output of the table and process it in Go to produce various metrics rather than try and do everything in SQL.

[Sticksman]

Should I close this one then?

[SuperQ]

@Sticksman Yes, I think the other metric is a superset of this. Do you agree?

[Sticksman]

Yep, sounds good. I'll close this one