Local Audit and Security Logs
Use Runtime’s audit records to trace a failed request, understand an access refusal, or inspect a lifecycle change. Records carry the owning operation’s result, correlation and time; an audit entry does not itself prove that a product journey succeeded.
What the Records Contain
Runtime records security-relevant authentication, authorization, execution, lifecycle and resource facts. Diagnostic projections are bounded and redacted: credentials, session proofs, raw private content and provider secrets do not belong in them.
Records are retained locally under count and encoded-size limits. Oldest records are evicted when those bounds are reached, and repeated identical refusals can share a counted record. Audit storage is not a permanent, exhaustive history of everything an App has done.
Investigate a Failed Request
- Keep the returned reason code, action hint and Trace ID when present.
- Use the available Runtime diagnostics to locate the matching operation and its actual outcome.
- For a Job, inspect the original Job ID before deciding whether another submission is safe. A missing audit row does not prove that an external action never happened.
Authorized audit export produces bounded diagnostic output; it does not imply automatic cloud aggregation or a Realm telemetry upload service. Handle exported records according to the information they actually contain.