We haven’t only been working on adding new regions for MPG, we’ve been quietly churning out some new features, too. The kind of things you’d expect from a managed service. We’ll get to them all in good time, but for now I’d like to draw your attention to the new Metrics tab.
We’ve kept things lean to start; you’ll be able to see:
- connections to Postgres and PgBouncer
- CPU and memory utilization for your DB and pooler instances
- database size over time
- database operations — selects/inserts/updates/deletes
- performance things like cache hit ratio and deadlocks
- replication lag/delay stuff
We’re exploring things like read-replicas, but for now replicas are for failover only—nonetheless we think it’s interesting for folks to know what is happening with replication.
We have more details about each of these charts in the docs: Monitoring and Metrics · Fly Docs
Clusters created before September 4th, 2025
If you have an older cluster, you’ll need to enable the new system to see the dashboard and for metrics to start flowing. This triggers a cluster restart, so you might see some client disconnections while that happens. Make sure you have appropriate reconnect logic in your app etc. etc.
Let us know what else you’d like to see. More metrics? Grafana integration? Forwarding to services like DataDog or New Relic? All possible!