On Sun, Feb 12, 2012 at 12:01 PM, Joshua Berkus <j...@agliodbs.com> wrote: > You'd pretty much need to do large-scale log harvesting combined with samples > of query concurrency taken several times per minute. Even that won't > "normalize" things the way you want, though, since all queries are not equal > in terms of the amount of data they hit. > > Given that, I'd personally take a statistical approach. Sample query > execution times across a large population of servers and over a moderate > amount of time. Then apply common tests of statistical significance. This > is why Heroku has the opportunity to do this in a way that smaller sites > could not; they have enough servers to (probably) cancel out any random > activity effects. >
Yes, I think if we could normalize, anonymize, and randomly EXPLAIN ANALYZE 0.1% of all queries that run on our platform we could look for bad choices by the planner. I think the potential here could be quite remarkable. -- Peter van Hardenberg San Francisco, California "Everything was beautiful, and nothing hurt." -- Kurt Vonnegut -- Sent via pgsql-performance mailing list (pgsql-performance@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-performance
