This has been a problem for a little while now and I still have no idea what's really going on - bodhi calls just seem to fail from time to time and at a rate much greater than what we've seen in the past. The problem is intermittent and even when we can reproduce it, there is no good data on what the root cause of the problem is - the client just indicates a http 500 regardless of where the problem actually was and debugging output sheds no additional light on the issue. At this time, we suspect network issues but that's not really enough to fix things.

As a workaround, implement a retry mechanism for read-only bodhi calls (list, mostly). When a read-only call fails (basically, everything other than comments), retry the call up to a configurable MAX_RETRIES value before throwing an exception.

The MAX_RETRIES value should have a sane default (3 sounds good) but needs to be configurable in taskotron.yml.