This page lists the actions required when upgrading JaaS. A release with no section here needs no migration — a plain helm upgrade (or new image tag) suffices.

2026.9.24

An operator that cannot reach the apiserver no longer ends the process. The pod stays Running and retries its manager in place, so the failures that used to show up as CrashLoopBackOff — egress denied to the apiserver, an incomplete ClusterRole, CRDs not installed — now show up as a pod that is Running, not Ready, and reconciling nothing. /jsonnet keeps serving throughout.

Alerting that watched container restarts for this class of failure goes quiet. Watch the operator’s own signals instead:

# The operator is not reconciling at all.
jaas_operator_available == 0

GET /operator on the management port carries the same state with the reason attached, and operator-unavailable covers diagnosis. Readiness behaves as before: a pod whose manager has never synced stays out of its Services, which still halts a rolling update.