I built our CI and tested failures and deployments. I never tested someone intentionally stopping CI while it was in progress. Two days after the team started u

I built our CI and tested failures and deployments. I never tested someone intentionally stopping CI while it was in progress. Two days after the team started using it, we realized we needed an immediate revert but the code was currently being built. So we just hit “stop” on the TC job. No big deal right? The current step was stopped, but it moved onto the next step like it passed. It deployed terribly broke things to production. This is when I learned TC sent SIGKILL on a stop and not SIGTERM, which my scripts were expecting.