From AWS RDS Performance Insights to Obsfly

• AWS-only. If you're multi-cloud or have on-prem DBs, you have a fork.
• No multi-host fleet view — every DB is its own console tab.
• No alerting. You have to glue CloudWatch alarms onto each metric.
• No AI advisory.

Fairness signal — useful in renewal conversations.

• Free for 7 days of retention on all RDS instances — useful as a baseline.
• Tight integration with RDS-native events.
• Top SQL view is reasonable for spot debugging.

1. Step 1

   Enable Obsfly for the same RDS instances

   Use the read-only monitoring user pattern documented in our RDS guide. No agent install required for RDS — Obsfly polls Performance Insights and pg_stat_statements directly.

2. Step 2

   Centralize alerting

   Replace per-instance CloudWatch alarms with Obsfly forecast-violation rules. One rule covers all hosts via tag selectors.

3. Step 3

   Add cross-fleet dashboards

   Obsfly's fleet status view shows all RDS hosts on one screen — color-coded by health, sized by QPS.

4. Step 4

   Keep PI enabled, optional

   PI is free; no need to disable. Use Obsfly as your primary tool, PI as a fallback.

• If you rely on CloudWatch alarms for autoscaling triggers, keep those — only migrate the alert definitions.
• RDS Proxy adds an extra hop; configure Obsfly to monitor both the proxy and the underlying instance.

Does Obsfly replace RDS Performance Insights?

Functionally yes — same metric coverage plus alerting, fleet view, anomaly detection, and AI insights. PI stays free for instances you haven't enrolled yet.

What about non-RDS Postgres?

Same agent, same UI. PI doesn't cover anything outside RDS — Obsfly does.