@map-colonies/prometheus

Overview

This package provides Express middleware for integrating Prometheus metrics into Node.js applications, making it easy to collect and expose application and request metrics.

Built on top of express-prom-bundle and prom-client, this package provides a simplified, opinionated configuration for MapColonies services.

Installation

npm install @map-colonies/prometheus prom-client
Copy

or

pnpm add @map-colonies/prometheus prom-client
Copy

Usage

Basic Setup

The simplest way to get started is to use both the metrics collection and endpoint middlewares:

import { collectMetricsExpressMiddleware, metricsMiddleware } from '@map-colonies/prometheus';
import express from 'express';
import { Registry } from 'prom-client';

const app = express();
const registry = new Registry();

// Collect metrics from all requests
app.use(collectMetricsExpressMiddleware({ registry }));

// Expose metrics at /metrics endpoint
app.get('/metrics', metricsMiddleware(registry));

// Your application routes
app.get('/', (req, res) => {
  res.json({ status: 'ok' });
});

app.listen(8080, () => console.log('Server listening on port 8080'));
Copy

Collect Metrics Middleware

The collectMetricsExpressMiddleware automatically measures request duration and counts, optionally collecting Node.js runtime metrics and service version information.

import { collectMetricsExpressMiddleware } from '@map-colonies/prometheus';
import express from 'express';
import { Registry } from 'prom-client';

const app = express();

const metricsMiddleware = collectMetricsExpressMiddleware({ 
  registry: new Registry(), 
  labels: { environment: 'production', team: 'platform' },
  collectNodeMetrics: true,
  collectServiceVersion: true
});

app.use(metricsMiddleware);
Copy

Metrics Endpoint Middleware

The metricsMiddleware provides a simple handler to expose collected metrics:

import { metricsMiddleware } from '@map-colonies/prometheus';
import { Registry } from 'prom-client';

const registry = new Registry();

// Use as a route handler (with default metrics collection)
app.get('/metrics', metricsMiddleware(registry));

// Or disable default metrics collection at the endpoint
app.get('/metrics', metricsMiddleware(registry, false));

// Or add labels to default metrics
app.get('/metrics', metricsMiddleware(registry, true, { environment: 'production' }));
Copy

Configuration Options

collectMetricsExpressMiddleware(options)

metricsMiddleware(registry, shouldCollectDefaultMetrics?, defaultMetricsLabels?)

Default Labels

The middleware automatically adds the following labels to all metrics:

• hostname: The system hostname (from os.hostname())
• service_name: The service name from package.json

Additional labels can be provided via the labels option.

Collected Metrics

The middleware automatically collects the following metrics:

Request Metrics

• http_request_duration_seconds (Histogram): Time taken to process requests, labeled by:
  • method: HTTP method (GET, POST, etc.)
  • status_code: HTTP response status code
  • operation: OpenAPI operation ID (if includeOperationId is enabled and express-openapi-validator is used)

Node.js Runtime Metrics

When collectNodeMetrics is enabled, the following metrics are collected:

• Process CPU usage
• Process memory usage (heap, external, RSS)
• Event loop lag
• Active handles and requests
• Garbage collection statistics

Service Information

When collectServiceVersion is enabled:

• service_version (Gauge): Service version information from package.json, labeled by:
  • service_version_major: Major version number
  • service_version_minor: Minor version number
  • service_version_patch: Patch version number
  • service_version_prerelease: Prerelease version (if applicable)

Advanced Examples

Custom Labels

import { collectMetricsExpressMiddleware } from '@map-colonies/prometheus';
import { Registry } from 'prom-client';

const registry = new Registry();

const middleware = collectMetricsExpressMiddleware({
  registry,
  labels: {
    environment: process.env.NODE_ENV || 'development',
    region: process.env.AWS_REGION || 'us-east-1',
    version: process.env.APP_VERSION || '1.0.0'
  }
});

app.use(middleware);
Copy

Transform Labels Based on Request

import { collectMetricsExpressMiddleware } from '@map-colonies/prometheus';
import { Registry } from 'prom-client';

const registry = new Registry();

const middleware = collectMetricsExpressMiddleware({
  registry,
  customLabels: {
    tenant: '',
    user_type: ''
  },
  transformLabels: (labels, req, res) => {
    // Add dynamic labels based on request
    labels.tenant = req.headers['x-tenant-id'] || 'unknown';
    labels.user_type = req.user?.type || 'anonymous';
    return labels;
  }
});

app.use(middleware);
Copy

Selective Metric Collection

import { collectMetricsExpressMiddleware } from '@map-colonies/prometheus';
import { Registry } from 'prom-client';

const registry = new Registry();

const middleware = collectMetricsExpressMiddleware({
  registry,
  collectNodeMetrics: true, 
  collectServiceVersion: true,
  includeOperationId: false // No OpenAPI validator
});

app.use(middleware);
Copy

Best Practices

1. Use a single registry: Share one Registry instance across your application to avoid duplicate metric collection.

2. Add meaningful labels: Use the labels option to add context-specific information (environment, region, etc.).

3. Avoid high-cardinality labels: Don't use user IDs, timestamps, or other high-cardinality values as labels, as this can cause memory issues.

4. Monitor metric cardinality: Regularly check the number of unique label combinations to prevent cardinality explosions.

Integration with Prometheus

To scrape metrics from your application, configure Prometheus with a scrape config:

scrape_configs:
  - job_name: 'my-nodejs-app'
    scrape_interval: 15s
    static_configs:
      - targets: ['localhost:8080']
    metrics_path: '/metrics'
Copy