fix(audit): retain batch on transient DB error and retry on next tick#272

Open

appleboy wants to merge 15 commits into

mainfrom

fix/audit-retry-on-transient-db-error

Commits on Jun 26, 2026

fix(audit): retain batch on transient DB error and retry on next tick

appleboy
and
claude
committed
fix(audit): report shutdown failure when final flush cannot persist records

appleboy
and
claude
committed
fix(audit): honour shutdown timeout on final flush; drop permanently-bad batches

appleboy
and
claude
committed
fix(audit): skip size-triggered flush while in retry state

appleboy
and
claude
committed
fix(audit): enforce maxRetryBuffer cap in addToBatch during retry state

appleboy
and
claude
committed
fix(audit): reset retry count on cap-eviction; report dropped records on shutdown

appleboy
and
claude
committed
fix(audit): remove maxFlushRetries drop; rely solely on maxRetryBuffer cap

appleboy
and
claude
committed
fix(audit): keep retry state after cap-eviction; add stale-tick record eviction

appleboy
and
claude
committed
fix(audit): set lastFlushDropped on all drop paths in flushBatchUnsafe

appleboy
and
claude
committed
fix(audit): remove staleTicks age-based eviction

appleboy
and
claude
committed
fix(audit): clamp ResourceName and Action to varchar(255) in buildAuditLog

appleboy
and
claude
committed
fix(audit): validate AuditDetails JSON serializability in buildAuditLog

appleboy
and
claude
committed
fix(audit): truncate ActorUserID to varchar(36) in buildAuditLog

appleboy
and
claude
committed
revert(audit): remove ActorUserID truncation

appleboy
and
claude
committed
fix(audit): truncate ActorUserID to varchar(36) with logged warning

appleboy
and
claude
committed