Monitoring detected elevated p99 latency on API endpoints (>500ms vs normal <100ms).

Root cause identified: database connection pool exhaustion due to a spike in concurrent Brand Audit requests.

Connection pool limits increased. Stale connections cleared.

Latency returned to normal levels. No data loss. All queued audit requests completed successfully.

Users reported failures when connecting TikTok accounts via OAuth flow.

Confirmed: TikTok Business API experiencing intermittent 503 errors (TikTok-side issue).

Implemented retry logic with exponential backoff for TikTok OAuth requests.

TikTok API restored. Backlogged connections completed. Added improved error messaging for future platform-side outages.

Planned maintenance window began. Application placed in read-only mode.

Database schema migration completed (new columns for Repurpose Engine moment detection scoring).

Data backfill completed. Integrity checks passed.

Full service restored. All systems operational.