My app newsbot-connectblue has had its machine replaced by your control plane approximately 2,755 times in the past 24 hours, creating a release roughly every minute without any action on my part. My app has been effectively unavailable. I am requesting immediate escalation, root cause explanation, and account credit. This began coinciding with the SIN distributed config store incident but has continued in IAD with no resolution.
Hi Bart, your machine is in IAD, it’s entirely unaffected by the SIN incident. The fly control plane is not creating these releases out of the blue, these happen when the app is deployed. Is it possible you have a cronned fly deploy or a periodic CI (e.g. github actions) looping deploys every minute or so?
The issue just stopped? Very strange. We didn’t change anything on our end.
How do you normally deploy this app? Do you have a CI process?
I’m curious if this was an actual fly infra problem or user error. If they were using GH to deploy, it may be stem from the GH issues that’s causing jobs to retrigger.
I don’t think it would be a Fly infra problem. What I see on my side are user-triggered deploys, like `fly deploy` was being run in a loop. It’s not the Fly built-in deployer (“deploy directly from github”). The most likely explanation is berserk CI or some kind of deploy-on-a-loop thing.