Question

I'm seeing repeated spikes to max CPU usage every 5 minute. The duration of the spikes is around 3 minutes.

I'm fairly certain that it's fogbugz, as I'm seeing a matching spike on the mysql server. I was wondering if it might be the reporting plugin, or something internal to fogbugz, as I don't have anything repeatedly hitting it other than Nagios, and that check is only looking for an open TCP port 80.

I'm very open to ideas, if you have them. I'm also including the usage chart, if this helps. Thanks!

Answer 1

The reporting/case history plugin or the search indexer can be fairly heavy hitters from time to time. Did you upgrade recently? It should level out eventually.

If you want to confirm, disabling the maintenance server temporarily will stop the above activities (along with mail import and other stuff you probably care about, so don't leave it disabled too long) and you can see if the spikes go away.

Answer 2

There's only three sources of load for FogBugz: the web interface, the XML API, and the maintenance service.

You say you shut down the maintenance service, so this load has to be coming from somewhere. Can you double-check that you're actually shutting down the maintenance service by shutting it down, then using the Send Email function?

If the mail gets sent, you might have another maintenance service out there running, which could itself be the problem.

If that's not it, then there might be some API script hitting the site. Tell your users they might be logged off, and if they are, to log back on with "Remember Me" not checked. Empty the Tokens table in the database, then wait ten minutes and see if a row appears in that table. If it does, then there's an API script running.