Operations

Recurring and incident tasks for a running installation, on Kubernetes or AWS Lambda. Day two answers "I need to" by task; the table below lists every page here.

When	Page
install, and check what you install	Install with Helm, deploy from a mirrored release, verify a release
the first day, then every day after	Day one, day two
somebody's access must end now	Revoke sessions
is it healthy	Check health, read the logs, read the audit trail, telemetry
keys and credentials	Rotate keys and credentials, rotate a client secret, sign with AWS KMS (deprecated)
backups	Back up and restore the installation, back up the legacy store and export
nobody can sign in	Lost operator access, recover on Lambda, turn off recovery
scale and availability	High availability, scaling and cache, provide a Valkey, a cold-start herd
storage and identity of the pod	Blobs on R2, switch the serving pod's role
upgrading	v1.77.3 and earlier pages, linked from the CHANGELOG