It looks like sea
is currently the leader. The check output there is stale, I don’t know why it’s not showing the role properly.
I think this failed because it stopped the primary first. sjc
didn’t get updated, so it couldn’t take over as primary. This is probably because it’s running multiple regions (which we love).
You probably need to do fly deploy --strategy immediate
to change the PRIMARY_REGION. This will replace all the VMs without waiting for health checks to pass. It will make the cluster inaccessible for a short time, though.