Fix: scale-to-zero silently broke the nightly backup - #10
Merged
Conversation
Today's scheduled backup failed with "app sentinel-license has no started VMs". Cause is the 2026-09-09 change that made this service scale to zero — my own change, and I did not think through this consequence. `flyctl ssh console` does not wake a stopped machine. Fly auto-starts on a connection to one of the app's SERVICES; SSH is not that path. So from the moment the app started sleeping, every scheduled run failed at the first step and no dump was written. Silent in the way that matters: nothing alerts on it. The only signal is a red X on a scheduled workflow nobody watches, and the failure mode of a backup is that you discover it at restore time. The job now starts the machine explicitly before SSH, then waits for the port to bind. Safe when already running, and the machine idles back to stopped on its own afterwards, so scale-to-zero is preserved. Note for anyone copying this workflow: Command Center's equivalent does NOT need this, because its web tier never sleeps. The two are otherwise meant to stay in step — this is a deliberate divergence, not drift. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Today's scheduled backup failed with
app sentinel-license has no started VMs.Cause is my own change — the 2026-09-09 move to scale-to-zero. I didn't think through this consequence.
flyctl ssh consoledoes not wake a stopped machine. Fly auto-starts on a connection to one of the app's services; SSH isn't that path. So from the moment this app started sleeping, every scheduled run failed at the first step and no dump was written.Silent in the way that matters: nothing alerts on it. The only signal is a red X on a scheduled workflow nobody watches — and the failure mode of a backup is that you find out at restore time.
The job now starts the machine explicitly before SSH, then waits for the port to bind. Safe when it's already running, and the machine idles back to stopped afterwards, so scale-to-zero is preserved.
Note for anyone copying this workflow: Command Center's equivalent does not need this, because its web tier never sleeps. The two are otherwise meant to stay in step — this is a deliberate divergence, not drift.
🤖 Generated with Claude Code