Home

Awesome

Roarr

NPM version Canonical Code Style Twitter Follow

JSON logger for Node.js and browser.

Motivation

For a long time I have been a big fan of using debug. debug is simple to use, works in Node.js and browser, does not require configuration and it is fast. However, problems arise when you need to parse logs. Anything but one-line text messages cannot be parsed in a safe way.

To log structured data, I have been using Winston and Bunyan. These packages are great for application-level logging. I have preferred Bunyan because of the Bunyan CLI program used to pretty-print logs. However, these packages require program-level configuration – when constructing an instance of a logger, you need to define the transport and the log-level. This makes them unsuitable for use in code designed to be consumed by other applications.

Then there is pino. pino is fast JSON logger, it has CLI program equivalent to Bunyan, it decouples transports, and it has sane default configuration. Unfortunately, you still need to instantiate logger instance at the application-level. This makes it more suitable for application-level logging just like Winston and Bunyan.

I needed a logger that:

In other words,

Roarr is this logger.

Usage

Producing logs

Roarr logger API for producing logs is the same in Node.js and browser.

  1. Import roarr
  2. Use any of the API methods to log messages.

Example:

import {
  Roarr as log,
} from 'roarr';

log('foo');

Consuming logs

Roarr logs are consumed differently in Node.js and browser.

Node.js

In Node.js, Roarr logging is disabled by default. To enable logging, you must start program with an environment variable ROARR_LOG set to true, e.g.

ROARR_LOG=true node ./index.js

All logs will be written to stdout.

Browser

In a browser, you must implement ROARR.write method to read logs, e.g.

import {
  ROARR,
} from 'roarr';

ROARR.write = () => {};

The API of the ROARR.write is:

(message: string) => void;

Example implementation:

import {
  ROARR,
} from 'roarr';

ROARR.write = (message) => {
  console.log(JSON.parse(message));
};

or if you are initializing ROARR.write before roarr is loaded:

// Ensure that `globalThis.ROARR` is configured.
const ROARR = globalThis.ROARR = globalThis.ROARR || {};

ROARR.write = (message) => {
  console.log(JSON.parse(message));
};

If your platform does not support globalThis, use globalthis polyfill.

You may also use @roarr/browser-log-writer that implements and opinionated browser logger with Liqe query support for filtering logs.

Filtering logs

Node.js

In Node.js, Roarr prints all or none logs (refer to the ROARR_LOG environment variable documentation).

Use @roarr/cli program to filter logs, e.g.

ROARR_LOG=true node ./index.js | roarr --filter 'context.logLevel:>30'

Browser

In a browser, Roarr calls globalThis.ROARR.write for every log message. Implement your own custom logic to filter logs, e.g.

globalThis.ROARR.write = (message) => {
  const payload = JSON.parse(message);

  if (payload.context.logLevel > 30) {
    console.log(payload);
  }
};

Log message format

Property nameContents
contextArbitrary, user-provided structured data. See context property names.
messageUser-provided message formatted using printf.
sequenceIncremental sequence ID (see adopt for description of the format and its meaning).
timeUnix timestamp in milliseconds.
versionRoarr log message format version.

Example:

{
  "context": {
    "application": "task-runner",
    "hostname": "curiosity.local",
    "instanceId": "01BVBK4ZJQ182ZWF6FK4EC8FEY",
    "taskId": 1
  },
  "message": "starting task ID 1",
  "sequence": "0",
  "time": 1506776210000,
  "version": "1.0.0"
}

API

roarr package exports a function with the following API:

export type Logger =
  (
    context: MessageContext,
    message: string,
    c?: SprintfArgument,
    d?: SprintfArgument,
    e?: SprintfArgument,
    f?: SprintfArgument,
    g?: SprintfArgument,
    h?: SprintfArgument,
    i?: SprintfArgument,
    k?: SprintfArgument
  ) => void |
  (
    message: string,
    b?: SprintfArgument,
    c?: SprintfArgument,
    d?: SprintfArgument,
    e?: SprintfArgument,
    f?: SprintfArgument,
    g?: SprintfArgument,
    h?: SprintfArgument,
    i?: SprintfArgument,
    k?: SprintfArgument
  ) => void;

To put it into words:

Refer to the Usage documentation for common usage examples.

adopt

<T>(routine: () => Promise<T>, context: MessageContext | TransformMessageFunction<MessageContext>) => Promise<T>,

adopt function uses Node.js async_context to pass-down context properties.

When using adopt, context properties will be added to all all Roarr messages within the same asynchronous context, e.g.

log.adopt(
  () => {
    log('foo 0');

    log.adopt(
      () => {
        log('foo 1');
      },
      {
        baz: 'baz 1',
      },
    );
  },
  {
    bar: 'bar 0',
  },
);
{"context":{"bar":"bar 0"},"message":"foo 0","sequence":"0","time":1506776210000,"version":"2.0.0"}
{"context":{"bar":"bar 0","baz":"baz 1"},"message":"foo 1","sequence":"0.0","time":1506776210000,"version":"2.0.0"}

sequence value

sequence represents async context hierarchy in ltree format, i.e.

<top-level sequential invocation ID>[.<async operation sequential invocation ID>]

Members of sequence value represent log index relative to the async execution context. This information can be used to establish the origin of the log invocation in an asynchronous context, e.g.

log.adopt(() => {
  log('foo 0');
  log.adopt(() => {
    log('bar 0');
    log.adopt(() => {
      log('baz 0');
      setTimeout(() => {
        log('baz 1');
      }, 10);
    });
    log('bar 1');
  });
});
{"context":{},"message":"foo 0","sequence":"0.0","time":1506776210000,"version":"2.0.0"}
{"context":{},"message":"bar 0","sequence":"0.1.0","time":1506776210000,"version":"2.0.0"}
{"context":{},"message":"baz 0","sequence":"0.1.1.0","time":1506776210000,"version":"2.0.0"}
{"context":{},"message":"bar 1","sequence":"0.1.2","time":1506776210000,"version":"2.0.0"}
{"context":{},"message":"baz 1","sequence":"0.1.1.1","time":1506776210010,"version":"2.0.0"}

Notice that even though logs baz 0 and baz 1 were produced at different times, you can tell that one was produced after another by looking at their sequence values 0.1.1.0 and 0.1.1.1.

Requirements

child

The child function has two signatures:

  1. Accepts an object.
  2. Accepts a function.

Object parameter

(context: MessageContext): Logger,

Creates a child logger that appends child context to every subsequent message.

Example:

import {
  Roarr as log,
} from 'roarr';

const barLog = log.child({
  foo: 'bar'
});

log.debug('foo 1');

barLog.debug('foo 2');
{"context":{"logLevel":20},"message":"foo 1","sequence":"0","time":1506776210000,"version":"2.0.0"}
{"context":{"foo":"bar","logLevel":20},"message":"foo 2","sequence":"1","time":1506776210000,"version":"2.0.0"}

Function parameter

<T>(context: TransformMessageFunction<MessageContext<T>>): Logger<T>

Creates a child logger that translates every subsequent message.

Example:

import {
  Roarr as log,
} from 'roarr';

const barLog = log.child<{error: Error}>((message) => {
  return {
    ...message,
    context: {
      ...message.context,
      ...message.context.error && {
        error: {
          message: message.context.error.message,
        },
      },
    },
  };
});

log.debug('foo 1');

barLog.debug({
  error: new Error('bar'),
}, 'foo 2');
{"context":{"logLevel":20},"message":"foo 1","sequence":"0","time":1506776210000,"version":"2.0.0"}
{"context":{"logLevel":20,"error":{"message":"bar"}},"message":"bar 2","sequence":"1","time":1506776210000,"version":"2.0.0"}

A typical use case for this pattern is serialization (e.g. of HTTP request, response or error object) and redaction of sensitive data from logs.

getContext

Returns the current context.

Example:

import {
  Roarr as log,
} from 'roarr';

const childLogger = log.child({
  foo: 'bar'
});

childLogger.getContext();

// {foo: 'bar'}

trace

debug

info

warn

error

fatal

Convenience methods for logging a message with logLevel context property value set to a numeric value representing the log level, e.g.

import {
  Roarr as log,
} from 'roarr';

log.trace('foo');
log.debug('foo');
log.info('foo');
log.warn('foo');
log.error('foo');
log.fatal('foo');

Produces output:

{"context":{"logLevel":10},"message":"foo","sequence":"0","time":1506776210000,"version":"2.0.0"}
{"context":{"logLevel":20},"message":"foo","sequence":"1","time":1506776210000,"version":"2.0.0"}
{"context":{"logLevel":30},"message":"foo","sequence":"2","time":1506776210000,"version":"2.0.0"}
{"context":{"logLevel":40},"message":"foo","sequence":"3","time":1506776210000,"version":"2.0.0"}
{"context":{"logLevel":50},"message":"foo","sequence":"4","time":1506776210000,"version":"2.0.0"}
{"context":{"logLevel":60},"message":"foo","sequence":"5","time":1506776210000,"version":"2.0.0"}

traceOnce

debugOnce

infoOnce

warnOnce

errorOnce

fatalOnce

Just like the regular logger methods, but logs the message only once.

Note: Internally, Roarr keeps a record of the last 1,000 Once invocations. If this buffer overflows, then the message is going to be logged again until the next time the buffer overflows again.

Utilities

getLogLevelName

Provides log level name (trace, debug, ...) for a numeric log level (10, 20, ...).

If numeric log level is between two ranges, then resolves to the one with greater severity (e.g. 5 => trace).

If numeric log level is greater than the maximum supported, then falls back to the greatest severity (fatal).

import {
  getLogLevelName,
} from 'roarr';
import type {
  LogLevelName,
} from 'roarr';

getLogLevelName(numericLogLevel: number): LogLevelName;

Middlewares

Roarr logger supports middlewares implemented as child message translate functions, e.g.

import {
  Roarr as log,
} from 'roarr';
import createSerializeErrorMiddleware from '@roarr/middleware-serialize-error';

const childLog = log.child(createSerializeErrorMiddleware());

const error = new Error('foo');

log.debug({error}, 'bar');
childLog.debug({error}, 'bar');
{"context":{"logLevel":20,"error":{}},"message":"bar","sequence":"0","time":1506776210000,"version":"2.0.0"}
{"context":{"logLevel":20,"error":{"name":"Error","message":"foo","stack":"[REDACTED]"}},"message":"bar","sequence":"1","time":1506776210000,"version":"2.0.0"}

Roarr middlewares enable translation of every bit of information that is used to construct a log message.

The following are the official middlewares:

Raise an issue to add your middleware of your own creation.

CLI program

Roarr CLI program provides ability to filter and pretty-print Roarr logs.

CLI output demo

CLI program has been moved to a separate package @roarr/cli.

npm install @roarr/cli -g

Explore all CLI commands and options using roarr --help or refer to @roarr/cli documentation.

Transports

A transport in most logging libraries is something that runs in-process to perform some operation with the finalized log line. For example, a transport might send the log line to a standard syslog server after processing the log line and reformatting it.

Roarr does not support in-process transports.

Roarr does not support in-process transports because Node processes are single threaded processes (ignoring some technical details). Given this restriction, Roarr purposefully offloads handling of the logs to external processes so that the threading capabilities of the OS can be used (or other CPUs).

Depending on your configuration, consider one of the following log transports:

Node.js environment variables

Use environment variables to control roarr behavior.

NameFunctionDefault
ROARR_LOGBooleanEnables/ disables logging.false
ROARR_STREAMSTDOUT, STDERRName of the stream where the logs will be written.STDOUT

When using ROARR_STREAM=STDERR, use 3>&1 1>&2 2>&3 3>&- to pipe stderr output.

Conventions

Context property names

Roarr does not have reserved context property names. However, I encourage use of the following conventions:

Context property nameUse case
applicationName of the application (do not use in code intended for distribution; see package property instead).
logLevelA numeric value indicating the log level. See API for the build-in loggers with a pre-set log-level.
namespaceNamespace within a package, e.g. function name. Treat the same way that you would construct namespaces when using the debug package.
packageName of the NPM package.

The roarr pretty-print CLI program is using the context property names suggested in the conventions to pretty-print the logs for the developer inspection purposes.

Log levels

The roarr pretty-print CLI program translates logLevel values to the following human-readable names:

logLevelHuman-readable name
10TRACE
20DEBUG
30INFO
40WARN
50ERROR
60FATAL

Using Roarr in an application

To avoid code duplication, you can use a singleton pattern to export a logger instance with predefined context properties (e.g. describing the application).

I recommend to create a file Logger.js in the project directory. Inside this file create and export a child instance of Roarr with context parameters describing the project and the script instance, e.g.

/**
 * @file Example contents of a Logger.js file.
 */

import {
  Roarr,
} from 'roarr';

export const Logger = Roarr.child({
  // .foo property is going to appear only in the logs that are created using
  // the current instance of a Roarr logger.
  foo: 'bar'
});

Roarr does not have reserved context property names. However, I encourage use of the conventions.

Recipes

Overriding message serializer

Roarr is opinionated about how it serializes (converts objects to JSON string) log messages, e.g. in Node.js it uses a schema based serializer, which is very fast, but does not allow custom top-level properties.

You can override this serializer by defining ROARR.serializeMessage:

import type {
  MessageSerializer,
} from 'roarr';

const ROARR = globalThis.ROARR = globalThis.ROARR || {};

const serializeMessage: MessageSerializer = (message) => {
  return JSON.stringify(message);
};

ROARR.serializeMessage = serializeMessage;

Logging errors

This is not specific to Roarr – this suggestion applies to any kind of logging.

If you want to include an instance of Error in the context, you must serialize the error.

The least-error prone way to do this is to use an existing library, e.g. serialize-error.

import {
  Roarr as log,
} from 'roarr';
import serializeError from 'serialize-error';

// [..]

send((error, result) => {
  if (error) {
    log.error({
      error: serializeError(error)
    }, 'message not sent due to a remote error');

    return;
  }

  // [..]
});

Without using serialization, your errors will be logged without the error name and stack trace.

Anti-patterns

Overriding globalThis.ROARR.write in Node.js

Overriding globalThis.ROARR.write in Node.js works the same way as it does in browser. However, overriding ROARR.write in Node.js is considered an anti-pattern because it defeats some of the major benefits outlined in Motivation section of the documentation. Namely, by overriding ROARR.write in Node.js you are adding blocking events to the event cycle and coupling application logic with log handling logic.

If you have a use case that asks for overriding ROARR.write in Node.js, then raise an issue to discuss your requirements.

Integrations

Using with Sentry

https://github.com/gajus/roarr-sentry

Using with Fastify

https://github.com/gajus/roarr-fastify

Using with Elasticsearch

If you are using Elasticsearch, you will want to create an index template.

The following serves as the ground work for the index template. It includes the main Roarr log message properties (context, message, time) and the context properties suggested in the conventions.

{
  "mappings": {
    "log_message": {
      "_source": {
        "enabled": true
      },
      "dynamic": "strict",
      "properties": {
        "context": {
          "dynamic": true,
          "properties": {
            "application": {
              "type": "keyword"
            },
            "hostname": {
              "type": "keyword"
            },
            "instanceId": {
              "type": "keyword"
            },
            "logLevel": {
              "type": "integer"
            },
            "namespace": {
              "type": "text"
            },
            "package": {
              "type": "text"
            }
          }
        },
        "message": {
          "type": "text"
        },
        "time": {
          "format": "epoch_millis",
          "type": "date"
        }
      }
    }
  },
  "template": "logstash-*"
}

Using with Scalyr

If you are using Scalyr, you will want to create a custom parser RoarrLogger:

{
  patterns: {
    tsPattern: "\\w{3},\\s\\d{2}\\s\\w{3}\\s\\d{4}\\s[\\d:]+",
    tsPattern_8601: "\\d{4}-\\d{2}-\\d{2}T[\\d:.]+Z"
  }
  formats: [
    {format: "${parse=json}$"},
    {format: ".*\"time\":$timestamp=number$,.*"},
    {format: "$timestamp=tsPattern$ GMT $detail$"},
    {format: "$timestamp=tsPattern_8601$ $detail$"}
  ]
}

and configure the individual programs to use RoarrLogger. In case of Kubernetes, this means adding a log.config.scalyr.com/attributes.parser: RoarrLogger annotation to the associated deployment, pod or container.

Using with NestJS

If you are using NestJS, you can use nestjs-logger-roarr to perform logging inside your application.

To enable one shared logger for the entire application:

import { RoarrLoggerService } from nestjs-logger-roarr';
import { AppModule } from "app.module";

const logger = RoarrLoggerService.sharedInstance();
const app = await NestFactory.create(AppModule, { logger });

To enable mutiple injected loggers via the usual Module syntax:

import { Module } from '@nestjs/common';
import { ConfigModule } from '@nestjs/config;
import { RoarrLoggerModule } from 'nestjs-logger-roarr';

@Module({
  imports: [
    RoarrLoggerModule.forRoot({
      logLevel: 'warn', // this is the *minimum* log level that will be displayed
    }),
  ]
})

export class AppModule {}

Documenting use of Roarr

If your package is using Roarr, include instructions in README.md describing how to enable logging, e.g.

## Logging

This project uses [`roarr`](https://www.npmjs.com/package/roarr) logger to log the program's state.

Export `ROARR_LOG=true` environment variable to enable log printing to `stdout`.

Use [`roarr-cli`](https://github.com/gajus/roarr-cli) program to pretty-print the logs.

Context Truncation

Roarr by default truncates context properties if the context object is wider or deeper than 10 properties. At the moment, this is a hard-coded value. Waiting for feedback on whether this is a reasonable default and if it needs to be configurable.

When the context goes over this limit, you will start seeing ... entries in your logs, e.g.

{"a":"a","b":"b","c":"c","d":"d","e":"e","f":"f","g":"g","h":"h","i":"i","j":"j","...":"1 item not stringified"}

The reason for this is to prevent accidental logging of massive objects that can cause context truncation and performance issues.

Developing

Every time a change is made to the logger, one must update ROARR_VERSION value in ./src/config.ts.

Unfortunately, this process cannot be automated because the version number is not known before semantic-version is called.