I am running a few web based applications with postgres on the backend. We have a few app servers load balanced all connecting to a dedicated postgres server. As usage on the applications increases I want to monitor my resources so that I can anticipate when I will hit bottlenecks on the db server. That way we can do upgrades or optimizations before our performance get's unacceptable.I use top to measure cpu and memory usage, vmstat to measure disk io, and pg_stat_activity to see which query is hogging the cpu, memory or io. We don't use raid so I'm not sure how vmstat deals with that.
We are monitoring cpu usage, memory usage, and network traffic. I would also like to monitor io utilization but am not quite sure how to do that. Does anyone here know to effetively monitor io to the raid (I'm guessing that this could be measured with the raid management software but was hoping their were some standard system commands.) or other io that I should be measuring?
What bottlenecks could I hit and what are some good stats to check to anticipate when I am moving towards one of those bottlenecks.
Thanks,
Rick
---------------------------(end of broadcast)--------------------------- TIP 5: Have you checked our extensive FAQ?
http://www.postgresql.org/docs/faqs/FAQ.html
So your mileage may vary :-)
Ron
---------------------------(end of broadcast)--------------------------- TIP 4: Don't 'kill -9' the postmaster