Metrics Library for Node.js
!npm packagenpm-imgnpm-url
!Downloadsdownloads-imgdownloads-url
!Issuesissues-imgissues-urlA simple, lightweight metrics library that provides an interface for instrumenting Node.js applications with Prometheus.
Motivation
This library is a JavaScript port of the metrics package written in Go lang by Tony Bellamy.Observability is a crucial aspect of building scalable and reliable applications. Instrumenting applications to gather meaningful metrics is a challenging task, especially when teams are not familiar with the fundamental concepts and best practices. This library aims to address these challenges by providing a streamlined approach to instrumenting Node.js applications with Prometheus. It draws inspiration from proven strategies and methodologies, such as the USE Method, RED Method, and The Four Golden Signals.
The primary goals of this library are to:
- Provide Minimal Viable Metrics (MVMs) based on well-established strategies, allowing teams to bootstrap their instrumentation efforts.
- Simplify the process of creating Prometheus counters, gauges, histograms, and summaries.
- Foster the adoption of Service Level Objectives (SLOs) and proven instrumentation practices.
Features
- Easy creation of Prometheus counters, gauges, histograms, and summaries.
- Implementation of MVMs based on the USE Method, RED Method, and The Four Golden Signals.
Naming
RED
class represents the RED method which focuses on measuring:
FourGoldenSignals
class represents The Four Golden Signals and focuses on measuring:
USE
class represents the USE method which focuses on measuring:
Usage
- Install the library
npm install node-key-metrics
- via yarn: yarn add node-key-metrics
- Import one of the MVM classes
import { USE } from 'node-key-metrics'
// Create an instance of USE with the desired configuration
// The DefaultMetric can be either collected or disabled based on the value of the collectDefaultMetrics parameter. By default disabled.
const useMetrics = new USE({
saturationName: 'saturation',
saturationHelp: 'saturation metric help text',
saturationLabels: ['label1', 'label2'],
utilizationName: 'utilization',
utilizationHelp: 'utilization metric help text',
utilizationLabels: ['label3', 'label4'],
collectDefaultMetrics: true,
})
// Increment the error counter
useMetrics.errors.inc()
// Set the saturation and utilization values
useMetrics.saturation.set({ label1: 'value1', label2: 'value2' }, 0.75)
useMetrics.utilization.set({ label3: 'value3', label4: 'value4' }, 0.5)
OR
- Use basic prometheus metrics:
import { counter, defaultMetrics, gauge, histogram, summary } from 'node-key-metrics'
// Register default metrics collector
defaultMetrics()
// Create a counter metric
const errorCounter = counter({
name: 'errors_total',
help: 'number of errors',
labelNames: ['error'],
})
// Increment the error counter
errorCounter.inc({ error: 'sample_error' })
// Create a gauge metric
const saturationGauge = gauge({
name: 'saturation',
help: 'saturation metric help text',
labelNames: ['label1', 'label2'],
})
// Set the saturation value
saturationGauge.set({ label1: 'value1', label2: 'value2' }, 0.75)
// Create a histogram metric
const responseTimeHistogram = histogram({
name: 'response_time',
help: 'response time metric help text',
labelNames: ['endpoint'],
})
// Observe a response time value
responseTimeHistogram.observe({ endpoint: '/api/some-endpoint' }, 100)
// Create a summary metric
const requestSizeSummary = summary({
name: 'request_size',
help: 'request size metric help text',
labelNames: ['endpoint'],
})
// Observe a request size value
requestSizeSummary.observe({ endpoint: '/api/another-endpoint' }, 1024)
- to expose the metrics to prometheus, use the
collect
method:
import { collect } from 'node-key-metrics'
app.get('/metrics', (req, res) => {
res.set('Content-Type', 'text/plain')
res.end(collect())
})