If swarming is a public API, in the sense that people and systems outside of chrome-infra's direct control can talk to machines that are also being used by chrome-infra's directly controlled machines, then, yes, we need some sort of quota or ddos mechanism.

buildbucket - which is also clearly a public API - also needs such a thing. 

Perhaps we need two bugs, then.

Issue 617846 is slightly related, as a recovery mechanism after the fact.

We care more about the workers, less about the server. Pending tasks can be canceled via scripts at worst, e.g. cancelling all state=PENDING, pool:Chrome-perf tasks:

./swarming.py query -S chromium-swarm.appspot.com --limit 0 'tasks/list?state=PENDING&tags=pool:Chrome-perf' > out.json
for task_id in $(python -c "import json; print '\n'.join(i['task_id'] for i in json.load(open('out.json'))['items'])"); do
  echo $task_id
  ./swarming.py cancel -S chromium-swarm.appspot.com  $task_id
done

This issue has been Available for over a year. If it's no longer important or seems unlikely to be fixed, please consider closing it out. If it is important, please re-triage the issue.

Sorry for the inconvenience if the bug really should have been left as Available. If you change it back, also remove the "Hotlist-Recharge-Cold" label.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot