Re: Rethinking stats communication mechanisms

Douglas McNaught <doug(at)mcnaught(dot)org> writes:

> (a) and (b): of course you would only do it on a temporary basis for
> problem diagnosis.

Temporary or not it isn't really an option when you're dealing with high
volumes. You could imagine a setup where say, 1% of page requests randomly
turn on debugging to get a random sample of database traffic. There are always
solutions.

But my point is that there's no reason to think that only long queries are
useful for snapshots. Short queries are equally capable of consuming resources
in aggregate. Hiding some subset of queries assuming they're uninteresting is
only going to produce deceptive results data.

> Ideally, you'd find the query storm problem in load testing before you
> ever got to production. I hope to someday visit that planet--it must
> be nice.

Indeed, when you get there send me a postcard :)

--
greg