Collect libp2p metrics for scraping by Prometheus or Graphana.

Example

import { prometheusMetrics } from '@libp2p/prometheus-metrics'

const metrics = prometheusMetrics()()
const myMetric = metrics.registerMetric({
 name: 'my_metric',
 label: 'my_label',
 help: 'my help text'
})

myMetric.update(1)

A metric that is expensive to calculate can be created by passing a calculate function that will only be invoked when metrics are being scraped:

Example

import { prometheusMetrics } from '@libp2p/prometheus-metrics'

const metrics = prometheusMetrics()()
const myMetric = metrics.registerMetric({
 name: 'my_metric',
 label: 'my_label',
 help: 'my help text',
 calculate: async () => {
  // do something expensive
   return 1
 }
})

If several metrics should be grouped together (e.g. for graphing purposes) registerMetricGroup can be used instead:

Example

import { prometheusMetrics } from '@libp2p/prometheus-metrics'

const metrics = prometheusMetrics()()
const myMetricGroup = metrics.registerMetricGroup({
 name: 'my_metric_group',
 label: 'my_label',
 help: 'my help text'
})

myMetricGroup.increment({ my_label: 'my_value' })

There are specific metric groups for tracking libp2p connections and streams:

Track a newly opened multiaddr connection:

Example

import { prometheusMetrics } from '@libp2p/prometheus-metrics'
import { createLibp2p } from 'libp2p'


const metrics = prometheusMetrics()()

const libp2p = await createLibp2p({
   metrics: metrics,
  })
// set up a multiaddr connection
const connection = await libp2p.dial('multiaddr')
const connections = metrics.trackMultiaddrConnection(connection)

Track a newly opened stream:

Example

import { prometheusMetrics } from '@libp2p/prometheus-metrics'
import { createLibp2p } from 'libp2p'

const metrics = prometheusMetrics()()

const libp2p = await createLibp2p({
  metrics: metrics,
})

const stream = await connection.newStream('/my/protocol')
const streams = metrics.trackProtocolStream(stream)