Dunno was the triaging process is, so handing it off directly.

Oh, the easiest way to answer this question is via a bot events query, here's an example:

./swarming.py query --limit 0 -S chromeos-swarming.appspot.com 'bot/chromeos-skylab-bot-06321e38-7463-4889-9752-4413899e9880/events?start=1541449500&end=1541535900'

That should give enough information to answer your question.

I think it still makes sense for this to be implemented Swarming side, since the client will need to make more queries to get the task details.  Either we query for each task individually after querying the task_complete events (which is expensive RPC-wise), or we still need to query all tasks sorted by created_ts, which gives us no guarantee how far back we need to query before we get all the tasks we want, especially as we start creating more low priority tasks that get run far after they were created to take advantage of spare capacity.

+1 to c#6, our experience with autotest suggests that trying to reorder list of tasks on client side doesn't scale well in our usecase because we have tasks that often run ~ 1 week after creation (by design: we have weekly low priority suites that slowly fill in the gaps between bursty latency sensitive tasks)