Mail Delivery Observability

Objective

Measure mail delivery health using message-level outcomes and destination patterns instead of noisy retry log volume.

Workflow

1. Identify authoritative mail logs, queue state and log rotation boundaries.
2. Correlate records by queue ID and define the observation window explicitly.
3. Separate unique messages from attempts/retries.
4. Calculate final/current outcomes: sent, deferred, bounced and unresolved where applicable.
5. Group failures by recipient domain, MX hostname/IP and normalized provider family when mapping is evidence-based.
6. Separate network/pre-banner failures from SMTP 4xx and 5xx responses.
7. Measure queue age and retry persistence so a small stuck cohort cannot masquerade as broad failure volume.
8. If multiple outbound transports exist, report direct vs relay volume and outcomes separately.
9. Alert on persistent patterns and changes in unique-message rates, not isolated retry lines.
10. Preserve enough timestamp/queue/MX evidence to support incident escalation without retaining unnecessary message content or recipient PII.