@samsystem/logger

Description

This plugin instantiates and exports an instance of winston.Logger configured to follow samsystems logging conventions for logging to kibana correctly. It also exports a logger factory function that can be used to explicitly pass options for creating winston.Logger instance as opposed to based on environment variables.

Getting started

• Clone repo: git clone https://github.com/samsystems/logger
• Run: npm install

Env vars example

Package is designed to run without any necessary setup. Just require('@samsystem/logger') and use in your service. It will log to kibana in beta/production environments and to console otherwise.

If you need to overwrite default behaviour you can use following env variables.

1. Disable all output

   LOG_SILENT=true

2. Set custom transports

   LOG_TRANSPORTS=kibana,console

3. Change log level

   LOG_LEVEL=info

4. MongoDB transport custom configuration

   LOG_MONGODB_URI=mongodb://localhost
   LOG_MONGODB_COLLECTION=logs
   LOG_MONGODB_LEVEL=error

5. File transport custom configuration

   LOG_FILE_FILENAME=output.log

6. Integrate with the debug Node package

   LOG_USE_DEBUG=1

   (and when invoking logger.debug)

   logger.debug('A useful log message', {
       trace: 'package:class:method',
   });

7. Splunk transport custom configuration

   LOG_SPLUNK_TOKEN=<token_access_splunk>
   LOG_SPLUNK_URL=<url_to_splunk>

Variables follow pattern:

1. General variables LOG_*
2. Transport specific variables LOGTRANSPORT* Transport specific variables overwrite general variables.

If you want to create custom logger you can use factory from lib/utils/createLogger.js to create your own logger.

How to use

Exported module is an instance of winston.Logger. If silent is false then the returned logger instance will use a instance of winston.transport.Console as the transport. Transports and logging-levels are able to be changed dynamically refer to docs

Using log-level and silent-mode set by environment variables

// pre-configured based on process.env.LOG_LEVEL  and process.env.LOG_SILENT
// defaults to {silent: false, level: 'info'}
const logger = require('@samsystem/logger');

logger.info('Inserted Customer', { customer: 'theCustomer' });
// {"customer":"theCustomer","level":"info","message":"Inserted Customer"}

logger.error('Error inserting customer', {
    error: 'the error message',
    customer: 'theCustomer',
});
// {"error":"the error message","customer":"theCustomer","level":"error","message":"Error inserting customer"}

logger.warn('Customer not found with email', { email: '[e-mail]' });
// {"email":"[e-mail]","level":"warn","message":"Customer not found with email"}

// LOGGING_LEVEL env variable must be set to debug in order for debug to show

logger.debug('the customer', { customer: 'theCustomer' });
// {"customer":"theCustomer","level":"debug","message":"the customer"}

Using createLogger method

const createLogger = require('@samsystem/logger/lib/utils/createLogger');
const logger = createLogger({
    // winston.Logger options
});

When using ElasticSearch transport

The ElasticSearch transport allows log data to be sent straight to ES without the need to be parsed by fluentd/logstash allowing for a more robust set of logs to be stored. In order to make it function properly the following environment variables need to be defined:

1. LOG_ELASTICSEARCH_INDEX This specifies which index to use, as a matter of best practice the index should be named after the application
2. LOG_ELASTICSEARCH_HOST The actual ES host to use, if the logger cannot connect to the ES host it will buffer a number of log messages (~5000) in memory and then bulk write them once it can heartbeat ElasticSearch

logger.tap

tap is a curried version of the logger, making it well suited for promise chains or functional composition.

Tap methods:

• tap
• tapInfo
• tapVerbose
• tapDebug
• tapWarn
• tapError
• tapCritical

You can call tap directly, and supply the log level you wish to use.

// tap
return getUser(username).then(
    logger.tap('debug', `Retrieved user with username: ${username}`),
);

Or you can use the convenience methods for each log level.

// tapInfo
return getUser(username).then(
    logger.tapInfo(`Retrieved user with username: ${username}`),
);

// tapVerbose
return getUser(username).then(
    logger.tapVerbose(`Retrieved user with username: ${username}`),
);

// tapDebug
return getUser(username).then(
    logger.tapDebug(`Retrieved user with username: ${username}`),
);

logger.tapAndThrow

tapAndThrow is similar to logger.tap, but instead of returning the logged value, it throws it.

Tap and Throw methods:

• tapAndThrow
• tapAndThrowInfo
• tapAndThrowVerbose
• tapAndThrowDebug
• tapAndThrowWarn
• tapAndThrowError
• tapAndThrowCritical

You can call tapAndThrow directly, and supply the log level you wish to use.

// tapAndThrow
return getUser(username)
    .then(logger.tapDebug(`Retrieved user with username: ${username}`))
    .catch(
        logger.tapAndThrow(
            'error',
            `Unexpected error finding user. username: ${username}`,
        ),
    );

Or you can use the convenience methods for each log level.

// tapAndThrowCritical
return getUser(username)
    .then(logger.tapDebug(`Retrieved user with username: ${username}`))
    .catch(
        logger.tapAndThrowCritical(
            `Unexpected error finding user. username: ${username}`,
        ),
    );

// tapAndThrowError
return getUser(username)
    .then(logger.tapDebug(`Retrieved user with username: ${username}`))
    .catch(
        logger.tapAndThrowError(
            `Unexpected error finding user. username: ${username}`,
        ),
    );

// tapAndThrowWarn
return getUser(username)
    .then(logger.tapDebug(`Retrieved user with username: ${username}`))
    .catch(
        logger.tapAndThrowWarn(
            `Unexpected error finding user. username: ${username}`,
        ),
    );