Recovery and escalation

Before correction

Record the time range, affected domains and mailboxes, message or job IDs, service state, latest errors and changes already made. When possible, create a support snapshot without message content.

Do not

• manually delete queue or migration records;
• modify audit tables;
• grant superuser or administrator rights as a temporary bypass;
• disable the Security Engine for an entire domain;
• restart every service without locating the cause;
• repeat an import over the same data without an idempotency check.

Escalation

Escalate when there is a risk of lost or duplicate mail, mailbox-store corruption, an unclear security decision, an incompatible database or schema version, or a post-update failure. Provide exact identifiers, reproducible steps and checks already completed.

After recovery, perform an end-to-end test: sign-in, internal message, inbound, outbound, trace, IMAP and worker heartbeat according to the affected scope.