While going through our uncategorized gate failures [1] I found that we have a lot of jobs failing (161 in 7 days) due to the tempest run timing out [2]. I originally thought it was just the networking scenario tests, but I was able to identify a handful of API tests that are also taking nearly 3 minutes each, which seems like they should be moved to scenario tests and/or marked slow so they can be run in a dedicated tempest-slow job. I'm not sure how to get the history on the longest-running tests on average to determine where to start drilling down on the worst offenders, but it seems like an audit is in order. [1] http://status.openstack.org/elastic-recheck/data/integrated_gate.html [2] https://bugs.launchpad.net/tempest/+bug/1783405 -- Thanks, Matt