Install

Requires Node.js 22 or newer. Works with ESM, CommonJS, and TypeScript.

npm install @sourceregistry/node-logger
Copy

Quick start

import {Console, LogLevel} from '@sourceregistry/node-logger';

const logger = Console(LogLevel.INFO);
logger.info('Server started', {port: 3000});
logger.withTags('api', 'orders').warn('Slow request', {durationMs: 850});
logger.error('Request failed', {requestId: 'abc'}, new Error('timeout'));
await logger.close();
Copy

The available levels are TRACE, DEBUG, INFO, WARN, ERROR, and FATAL. Each logging method returns immediately. Call await logger.flush() when you need queued writes to finish, and await logger.close() during shutdown. Transport failures surface from these calls as an AggregateError.

For commands and required environment variables, see the runnable examples.

Transports

Choose a transport for each destination. You can combine several in one logger:

Transport	Runnable example
Console	Console
File	File
Buffered file	Buffered file
Smart file	Smart file
HTTP	HTTP
Elasticsearch	Elasticsearch
Worker	Worker

import {ConsoleTransport, FileTransport, JSONFormatter, Logger, LogLevel} from '@sourceregistry/node-logger';

const logger = new Logger(LogLevel.INFO, [
  new ConsoleTransport(),
  new FileTransport('./logs/app.jsonl', new JSONFormatter()),
]);
logger.info('Order accepted', {orderId: '123'});
await logger.close();
Copy

Console

Console() is a shortcut for a text formatted console logger. To choose a formatter, create the transport directly:

import {ConsoleTransport, JSONFormatter, Logger, LogLevel} from '@sourceregistry/node-logger';

const logger = new Logger(LogLevel.INFO, [new ConsoleTransport(new JSONFormatter())]);
logger.info('Ready', {port: 3000});
await logger.close();
Copy

File

File() writes human readable logs to the console and JSON Lines to a file:

import {File, LogLevel} from '@sourceregistry/node-logger';

const logger = File('./logs/app.jsonl', LogLevel.INFO);
logger.info('Payment received', {paymentId: 'pay_123'});
await logger.close();
Copy

Use FileTransport directly when you only want file output, as in the combined example above.

Buffered file

Buffer entries until the batch reaches 100 entries or one second passes:

import {BufferedFileTransport, JSONFormatter, Logger, LogLevel} from '@sourceregistry/node-logger';

const logger = new Logger(LogLevel.INFO, [
  new BufferedFileTransport('./logs/events.jsonl', new JSONFormatter(), LogLevel.INFO, 100, 1_000),
]);
logger.info('Event processed', {eventId: 'evt_123'});
await logger.close();
Copy

Smart file

Flush by interval, batch size, severity, or idle time:

import {JSONFormatter, Logger, LogLevel, SmartFileTransport} from '@sourceregistry/node-logger';

const logger = new Logger(LogLevel.INFO, [
  new SmartFileTransport('./logs/app.jsonl', new JSONFormatter(), LogLevel.INFO, {
    enabled: true,
    interval: 5_000,
    onSize: 50,
    onLevel: LogLevel.ERROR,
    onIdle: 10_000,
  }),
]);
logger.error('Database unavailable');
await logger.close();
Copy

Buffered file transports keep a bounded in-memory queue. close() flushes remaining entries. A completed stream write does not guarantee durability after a power loss; FileTransport.flush() requests fsync where a file descriptor is available.

HTTP

Send batches to an HTTP endpoint:

import {HTTPTransport, Logger, LogLevel} from '@sourceregistry/node-logger';

const logger = new Logger(LogLevel.INFO, [
  new HTTPTransport({
    endpoint: 'https://logs.example.com/ingest',
    headers: {Authorization: `Bearer ${process.env.LOG_TOKEN}`},
    batchSize: 50,
    maxQueueSize: 10_000,
    requestTimeout: 10_000,
  }),
]);
logger.info('Job finished', {jobId: 'job_123'});
await logger.close();
Copy

The endpoint receives a JSON array of formatted strings. The transport retries network failures, HTTP 408/429, and HTTP 5xx responses. A failed batch is reported by flush() or close() and is not automatically requeued because delivery may have succeeded before the connection failed.

Set LOG_TOKEN in the application environment before running this example.

Elasticsearch

Send bulk NDJSON to an Elasticsearch /_bulk endpoint:

import {ElasticsearchTransport, Logger, LogLevel} from '@sourceregistry/node-logger';

const logger = new Logger(LogLevel.INFO, [
  new ElasticsearchTransport({
    endpoint: 'https://es.example.com/_bulk',
    index: 'application-logs',
    apiKey: process.env.ELASTIC_API_KEY,
  }),
]);
logger.info('Document indexed', {documentId: 'doc_123'});
await logger.close();
Copy

The transport checks the bulk response for item errors and reports them on flush() or close().

Set ELASTIC_API_KEY if your cluster requires API key authentication.

Worker

Save this worker as logger-worker.mjs:

import {parentPort} from 'node:worker_threads';

parentPort.on('message', message => {
  if (message.type === 'log') process.stdout.write(`${message.data}\n`);
});
Copy

Then connect it to the logger:

import {resolve} from 'node:path';
import {Logger, LogLevel, WorkerTransport} from '@sourceregistry/node-logger';

const logger = new Logger(LogLevel.INFO, [
  new WorkerTransport(resolve('logger-worker.mjs')),
]);
logger.info('Background task started', {taskId: 'task_123'});
process.once('SIGINT', () => { void logger.close().catch(console.error); });
Copy

The worker keeps this example running until Ctrl+C. Worker delivery is best effort: messages already posted to a worker are not acknowledged, and close() terminates the worker. Use a file or HTTP transport when confirmed delivery at shutdown matters.

Formatters

Pass a formatter to a transport to control its output. These examples write to the console; the same formatters work with file, HTTP, and worker transports.

Formatter	Runnable example
JSON	JSON
Text	Text
CEF	CEF
Syslog	Syslog

JSON

import {ConsoleTransport, JSONFormatter, Logger, LogLevel} from '@sourceregistry/node-logger';

const logger = new Logger(LogLevel.INFO, [new ConsoleTransport(new JSONFormatter())]);
logger.info('User signed in', {userId: 'user_123'});
await logger.close();
Copy

JSONFormatter includes timestamps, levels, context, tags, and errors. It converts bigint values to strings and replaces circular references with [Circular].

Text

import {ConsoleTransport, Logger, LogLevel, TextFormatter} from '@sourceregistry/node-logger';

const logger = new Logger(LogLevel.INFO, [
  new ConsoleTransport(new TextFormatter(true, false)), // timestamp, no ANSI colors
]);
logger.info('Listening', {port: 3000});
await logger.close();
Copy

CEF

import {CEFFormatter, ConsoleTransport, Logger, LogLevel} from '@sourceregistry/node-logger';

const logger = new Logger(LogLevel.INFO, [
  new ConsoleTransport(new CEFFormatter('Acme', 'API', '1.0')),
]);
logger.warn('Access denied', {requestId: 'req_123'});
await logger.close();
Copy

Syslog

import {ConsoleTransport, Logger, LogLevel, SyslogFormatter} from '@sourceregistry/node-logger';

const logger = new Logger(LogLevel.INFO, [
  new ConsoleTransport(new SyslogFormatter(16, 'api-1', 'orders-service')),
]);
logger.error('Dependency unavailable', {service: 'database'});
await logger.close();
Copy

SyslogFormatter produces RFC 5424 style records. Your log collector still needs a transport that sends records to it.

Shutdown and errors

process.once('SIGTERM', () => {
  void logger.close().then(
    () => process.exit(0),
    error => {
      console.error('Log delivery failed during shutdown:', error);
      process.exit(1);
    },
  );
});
Copy

close() is safe to call more than once. Logging after it starts throws. The logger continues sending an entry to other transports if one fails; errors are reported when you flush or close it. Context is not redacted automatically, so remove secrets and personal data before logging.

For custom destinations, implement Transport with write(entry) and optionally flush() and close().

License

Apache-2.0. SourceRegistry logo from the SourceRegistry organization, licensed CC BY 4.0.