// the find
phoenixframework/phoenix_live_dashboard
Realtime dashboard with metrics, request logging, plus storage, OS and VM insights
LiveDashboard is the built-in admin panel that ships with Phoenix — real-time visibility into your running Erlang VM, OS resources, Ecto query stats, and live telemetry metrics, all over WebSockets via LiveView. It's for Phoenix developers who want production-grade observability without standing up a separate tool. Ships by default in new Phoenix apps, so most Elixir developers already have it.
The multi-node support is genuinely useful — you can inspect a remote BEAM node from your dashboard node without extra tooling, which is rare in web frameworks. The Ecto stats page shows actual index and table usage from pg_stat, not just query counts, so you can spot missing indexes without leaving the dashboard. The `PageBuilder` behaviour lets you mount custom pages, so it's extensible rather than a closed black box. Process and supervision tree inspection is a direct window into the BEAM runtime in a way no other language's tooling can match.
2156 stars is surprisingly low for something that ships with every Phoenix app — it's underrated but also a signal that most teams treat it as a dev-only tool and never bother securing it for production, which is a footgun the README only half-addresses. Historical metrics require you to wire up your own storage (metrics_history guide points to an external library); out of the box everything is in-memory and lost on restart, so it's useless for post-incident analysis. The request logger works by subscribing to a PubSub topic, which means in a multi-node cluster you only see requests from the node you're connected to — easy to miss this until you're chasing a bug on node B. The UI is functional but dated; charts are basic and there's no alerting or threshold configuration.