963128 - Enable ANALYZE_MODEL_FETCHES on crash stats dev and stage

Status: RESOLVED FIXED

(Infrastructure & Operations Graveyard :: WebOps: Socorro, task)

Description

[Peter Bengtsson [:peterbe]]

As of https://bugzilla.mozilla.org/show_bug.cgi?id=963126 we have a way of measuring how we're using caches when talking to the middleware. 

This new measuring technique thing is, by default, disabled and we should probably test it out properly first on dev and stage.

Can you set `ANALYZE_MODEL_FETCHES = True` in settings/local.py for dev and stage?

Comment 1

[Brandon Burton [:solarce]]

(In reply to Peter Bengtsson [:peterbe] from comment #0)
> As of https://bugzilla.mozilla.org/show_bug.cgi?id=963126 we have a way of
> measuring how we're using caches when talking to the middleware. 
> 
> This new measuring technique thing is, by default, disabled and we should
> probably test it out properly first on dev and stage.
> 
> Can you set `ANALYZE_MODEL_FETCHES = True` in settings/local.py for dev and
> stage?

-dev is being retired, would you still like this for staging?

Comment 2

[Peter Bengtsson [:peterbe]]

Yes please. 
It's still a bit weird because stage doesn't get nearly enough traffic to be realistic but it'll test that the new measurement code really works. And when we know it works, we can enable it on prod for a week or something.

Comment 3

[Brandon Burton [:solarce]]

This has been added to the local.py template for staging and been pushed live via puppet, apache has been restarted

Sending        socorro/templates/stage/etc-socorro/local.py.erb
Transmitting file data .
Committed revision 82247.