Pino Logger Config Builder

Pino is a structured JSON logger for Node.js, widely regarded as one of the fastest loggers available. Its speed comes from minimizing object allocation during log calls, using a worker thread (asynchronous) for I/O when possible, and serializing log records to JSON strings as efficiently as possible. Unlike Winston, Pino separates concerns — the logger writes to stdout and a separate process (pino-pretty, log rotation, cloud transport) handles formatting.

pino-pretty is a development-time formatter that converts Pino's compact JSON output into a human-readable colorized format. Install it as a dev dependency: npm install -D pino-pretty. Use it only in development — in production, keep raw JSON output which is faster and directly ingestable by log aggregators (Datadog, Loki, CloudWatch). Configure it as a transport: { target: 'pino-pretty', options: { colorize: true } }.

For Fastify: pass logger: true (or logger: { level: 'info' }) in the Fastify constructor — Fastify uses Pino internally. For Express: import pino and pino-http, then use app.use(pinoHttp({ logger: pinoInstance })). The pinoHttp middleware automatically logs each request/response pair including status code, response time, and content length.

redact removes or masks sensitive fields from log output. Pass redact: ['password', 'req.headers.authorization', 'user.email'] to replace those paths with '[Redacted]'. Use 'req.headers.authorization' to hide Bearer tokens, '*.password' to redact passwords at any depth. Redaction uses fast-redact under the hood and runs at serialization time with near-zero overhead.

const child = logger.child({ requestId: '123' }) creates a new logger that inherits the parent's config but adds fixed fields to every log line. This is the idiomatic way to add context — in web servers, create a child logger per request with { requestId, userId } so every log from that request automatically includes that context without passing it manually to every function.

Pino has 6 levels in ascending severity: trace (5) for verbose debugging, debug (10) for development diagnostics, info (20) for normal operation events, warn (30) for degraded but functioning states, error (40) for failures that need attention, fatal (50) for critical failures that crash the app. Setting level: 'info' suppresses trace and debug. In production, use 'info' or 'warn'. In development, use 'debug' or 'trace'.