Clearinghouse SLA

Clearinghouse SLA, at /admin/clearinghouse/sla, gives a super-admin a real-time read on how well cross-product event delivery is performing. A toolbar picks the time window (1h, 24h, 7d, 30d, default 24h) and can filter by tenant, though the registry-based metrics are global. Five panels report the health signals: Projection latency shows p50, p95, and p99 with a sample count, turning red when p95 reaches 1000 ms; Error rate shows failed over total, red at 5%; Throughput shows events per minute over the window; Escalation rate shows open escalations over total registry rows, red at 10%; and Audit drain backlog combines three signals — the age of the oldest undrained audit-outbox row (red at 1 hour), the count of dead-lettered rows, and any recent audit_drain_failed alerts. Together these panels tell operators whether events are flowing on time and whether the audit trail is keeping up.