Just deployed v2.4 to production. All health checks are green.
The new rate limiter is active on the API gateway. 1000 req/min per user.
Nice, the dashboard latency already dropped. Seeing p99 at 120ms down from 340ms.
The Sentry error rate is way down too. Only 2 errors in the last hour vs 47 yesterday.
Both were timeout retries that resolved on the second attempt.
Can someone review my PR for the notification service? It adds WebSocket support for real-time alerts.
I'll take a look after lunch. Is it based on the RFC we discussed last sprint?
Yep, follows the spec exactly. Added reconnection logic with exponential backoff and a heartbeat every 30s.
Heads up: I'm rotating the database credentials at 2 PM UTC. There will be a brief connection reset.
The connection pool should handle it automatically but keep an eye on your logs.
Thanks for the heads up. I'll set a Grafana alert for connection errors during that window.