One App includes a metrics server which exposes a set of metrics using the prom-client
which can be easily scraped by a service such as Prometheus.
Metrics will be exposed on the port configured by the environment variable HTTP_METRICS_PORT.
They can be collected from the /metrics path, for example http://localhost:3005/metrics.
HTTP_METRICS_PORT defaults to
3005.
One App makes the following metrics available:
- One App Version
- Circuit breaker, performance including open, close events,
- Module Map, updates, restarts and errors
- CPU
- Memory
- Garbage collection
- Eventloop lag
If you wish to use Grafana the one-app-classic dashboard can provide you with a good starting point.
It possible to run Prometheus and Grafana against a locally running build of one-app. While not an ideal performance indicator you could use it to gauge the impact your module might have.
Run the following inside of One App to start the performance monitoring locally:
npm run perf -- monitor
There is not an option to run performance monitoring with
one-app-runner, though instances ofone-app-runnercan be monitored with the--targetoption.
More details for running performance monitoring can be found in the performance readme/