Instance startup fails when ENCORE_RUNTIME_CONFIG secret version is destroyed during autoscaling #2202
Description
Just to note: this isn't urgent as a redeploy resolved the issue. Wanted to share in case it's helpful for others or indicates something worth investigating on the platform side
Description
When we deployed, instances fail to start because they're trying to fetch a destroyed secret version for ENCORE_RUNTIME_CONFIG.
Error Logs
Starting new instance. Reason: AUTOSCALING - Instance started due to configured scaling factors
Could not fetch secret "projects/ec-prod-apprun-euw1d/secrets/apprun-XXXX-ENCORE_RUNTIME_CONFIG/versions/243" for environment variable "ENCORE_RUNTIME_CONFIG". Instance startup will now abort: Secret Version [...] is in DESTROYED state.
The request failed because the instance could not start successfully.
Expected Behaviour
Autoscaling instances should be able to start successfully and fetch the current valid secret version.
Actual Behaviour
New instances attempt to fetch a destroyed secret version and crash, causing request failures.
Workaround
Redeploying the application (encore deploy --env prod) creates a new secret version and resolves the issue.
Environment
- Platform: Encore Cloud (GCP Cloud Run)
- Environment: Production