// the take
PromEx wires Telemetry events from Phoenix, Ecto, Oban, LiveView, Absinthe, and a handful of other BEAM libraries into Prometheus metrics, and ships matching Grafana dashboards so you're not building panels from scratch. It's for Elixir teams already running Prometheus/Grafana who want library-level metrics without writing their own Telemetry handlers.
The plugin behaviour is a clean abstraction, and each plugin comes with its own tested Grafana dashboard JSON generated via EEx templates, so metrics and visualization ship together instead of as an afterthought. It has a real test suite per plugin, a config renderer, and a dashboard uploader/linter with their own tests, not just a thin wrapper around telemetry_metrics. It's opinionated by design: metric names and tags aren't configurable through options, which keeps the dashboards from rotting when someone tweaks a label — if you disagree, you fork the plugin, and the maintainer says as much upfront.
Coverage for anything outside the Phoenix/Ecto/Oban stack is thin — Finch and Redix have been listed as 'Coming soon' across multiple releases, so if your app leans on either you're writing your own plugin. It bundles a Grafana Agent binary downloader (lib/prom_ex/grafana_agent/downloader.ex) that pulls an external binary at setup time, which is an extra supply-chain surface most metrics libraries don't introduce. The /metrics endpoint has zero auth by default and the README is explicit that securing it is entirely on you — easy to forget in a public-facing deploy. It's effectively a one-maintainer project (akoutmos, who also pitches paid support in the README), so plugin requests for less common libraries move at the pace of one person's bandwidth.
We read the README, the directory tree and a key source file in September 2026. Nobody pays to be listed. How we judge →