tflink

#!/usr/bin/env python
""" Collect data on fedmsg reliability.

The gist is that we start up and go into a loop.  Each time we wake up we:
    - Ask koji for all the builds that were started in the last hour.  This
      includes builds that eventually fail or are in progress.
    - Ask fedmsg for all the koji events from the past error.  Throw out all
      the ones except the 'build start' events.
    - Compare the number of new koji builds to the number of fedmsg new build
      messages.  They should be the same.