.. | ||
autonat | ||
autorelay | ||
eventbus | ||
holepunch | ||
identify | ||
relaysvc | ||
resource-manager | ||
swarm | ||
dashboard.yml | ||
datasources.yml | ||
docker-compose-linux.yml | ||
docker-compose.base.yml | ||
prometheus.yml | ||
README.md |
Grafana Dashboards
This directory contains prebuilt dashboards (provided as JSON files) for various components. For steps on how to import and use them please read the official Grafana documentation.
Public dashboards
You can check the following prebuilt dashboards in action:
These metrics come from one of the public IPFS DHT bootstrap nodes run by Protocol Labs. At the time of writing (2023-08), these nodes handle many connections across various libp2p implementations, versions, and configurations (they don't handle large file transfers).
Using locally
For local development and debugging, it can be useful to spin up a local Prometheus and Grafana instance.
To expose metrics, we first need to expose a metrics collection endpoint. Add this to your code:
import "github.com/prometheus/client_golang/prometheus/promhttp"
go func() {
http.Handle("/debug/metrics/prometheus", promhttp.Handler())
log.Fatal(http.ListenAndServe(":5001", nil))
}()
This exposes a metrics collection endpoint at http://localhost:5001/debug/metrics/prometheus. Note that this is the same endpoint that Kubo uses, so if you want to gather metrics from Kubo, you can skip this step.
On macOS:
docker compose -f docker-compose.base.yml up
On Linux, dashboards can be inspected locally by running:
docker compose -f docker-compose.base.yml -f docker-compose-linux.yml up
and opening Grafana at http://localhost:3000.
Making Dashboards usable with Provisioning
The following section is only relevant for creators of dashboards.
Due to a bug in Grafana, it's not possible to provision dashboards shared for external use directly. We need to apply the workaround described in https://github.com/grafana/grafana/issues/10786#issuecomment-568788499 (adding a few lines in the dashboard JSON file).