Materialize/misc/monitoring/README.md

Local monitoring composition

An mzcompose composition for running a minimal monitoring stack of Prometheus (metrics), Tempo (distributed tracing), and Grafana locally. Metrics from all processes run by the bin/environmentd script will be automatically discovered and collected, and traces will be collected when run with --monitoring:

./bin/environmentd --monitoring

Usage

cd misc/monitoring
./mzcompose run default

To access Grafana, run ./mzcompose web grafana or navigate directly to http://localhost:3000 in your browser. Grafana will have datasources for both Prometheus and Tempo by default.

If needed, to access Prometheus directly, run ./mzcompose web prometheus or navigate directly to http://localhost:9090 in your browser.

Tempo does not have its own UI and can be only accessed via Grafana. See the tracing docs for more info on how to filter and access tracing data.

Modifications

You can adjust the Prometheus and Tempo configurations by editing the prometheus.yml and tempo.yml files in this directory. They are bind mounted into the containers, so edits will be picked up the next time you restart:

./mzcompose down -v
./mzcompose run default

You can adjust the Grafana configuration through its UI. Beware that the Grafana state is ephemeral, and will be lost the next time you run ./mzcompose down.