Incident communication

If you experience a service disruption, record the affected endpoint or product, first observed time, region, request identifiers that are safe to share, and whether retries change the result. Do not send secrets or full customer payloads. The support team uses the available service and operational signals to investigate, provides updates when material information changes, and records the customer-visible outcome. During an incident, avoid repeated destructive retries and keep a small timestamped sample of the observed behaviour. When service is restored, confirm whether your workload needs a controlled retry or reconciliation. A support ticket can link the incident to your account and preserve the evidence needed for follow-up.