for people running cnpg in prod, what caught you out that you didnt see coming?

one thing we noticed: a cluster with no backups configured at all still shows ContinuousArchiving=True, "Continuous archiving is working". the archive command just returns ok when theres nowhere to send the wal. kubectl cnpg status does say "Continuous Backup not configured", but an alert on the condition alone would miss it

curious about the rest. failovers, wal filling the disk, upgrades, restores that didnt work, whatever it was, and how you found out

full disclosure i work on radar (open source k8s ui). we've been gradually adding cnpg support and want to make sure it covers the stuff that actually bites people