Fixed :)
Our server restarted recently (the host had a failure on the node we are hosting on) and the frontends didn't come back cleanly it seems. As we recently changed our endpoint monitoring (new status page: https://status.reddthat.com) I was only doing the
I've modified the all the frontends such that:
- deployments will come back correctly now after an (unexpected) server restart
- All alternate-frontends are now increased from 1 to 2, so we have a little "HA" incase one container crashes
- I've added monitoring to the frontends
That should sort out the frontends so they are no longer "second class citizens".