Our app is completely stuck.
The website gives 502.
From the logs we see we have an old release that is running on a VM, but this VM gives tons of errors.
We launched a new deploy, new VMs gets created, but the old ones are not stopped (including the one that returns a lot of errors).
Now in the dashboard we see 6 VMs instead of 2. But we cannot see the one that is giving the errors.
Also, we tried to stop that VM, but we get this error:
Error failed to stop allocation: Could not find Allocation
We cannot stop the VMs, nor restart, …
Is it a regular multi-process Fly app? I’ve seen similar (but less severe) issues with those once before.
See if these steps give you any hints on what’s going on:
In the meanwhile, to mitigate the downtime, you can deploy the same code to a new Fly application and flip your DNS records to point to those. A caveat, keep this recent shared-IPv4 related change in mind (just in case your Fly app expects dedicated IPv4s): Announcement: Shared Anycast IPv4