Can I fail a "guided failure mode" deploy after x amount of time?

Badly worded title, so let me elaborate. We have a few projects that need guided failure mode, however sometimes users will leave a deploy overnight, which can cause hang-ups on other projects due to limited deployment targets/deployment concurrency. Not all users have access to all projects, so if a guided failure is left unaddressed in project 1, a user with access to only project 2 cannot fail that deploy in order to run their own. Across global timezones, this is tricky. We’ve tried education and documentation, but the issue persists, so:

Is it possible to automatically cancel a “forgotten” guided failure deployment after x amount of time?

Hi @jeffrey.mitchell,
You could possibly do something with a scheduled runbook to call the API and check for “forgotten” guided failure deployments, but I’m wondering if they are being forgotten or left for an extremely long time is guided failure as the default the best way?

Another option is selecting Guided Failure for that specific deployment if they are going to be around to monitor it and handle any failures.

If they are kicking off deployments and walking away, would it be better to just let them fail and then research the issue and redeploy?

Mark Reeder

Hey Mark,

Definitely don’t disagree on taking guided failure off those deploys. Some of them are required as such though, unfortunately a decision out of my hands. We could definitely go the education route of “only turn on guided failure if you plan on guiding it” but yeah, could look into the runbook. Appreciate the tips!