Worker nodes don't automatically reconnect after failover - restart required

When performing a failover process for Spacelift self-hosted upgrades, worker nodes will not automatically reattach to the secondary instance even after updating the DNS record for the domain. This is expected behavior.

Why this happens

Worker nodes establish their connection to the Spacelift backend during instance startup and maintain that persistent connection throughout their lifecycle. When you failover to a secondary instance and update the DNS record, the existing worker nodes continue using their original connection and don't automatically detect or switch to the new instance.

Required action after failover

To complete the failover process, you need to restart all worker nodes so they can establish new connections to the secondary instance. This can be done by:

1. Setting the desired count of worker instances to 0

2. Waiting for all instances to terminate

3. Resetting the desired capacity to your original configuration

The new worker instances will then start up and automatically connect to the secondary Spacelift instance using the updated DNS record.