No release in over 3 years
Low commit activity in last 3 years
Ruby logger that writes directly to LogStash
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
 Dependencies

Development

Runtime

~> 1.0
 Project Readme

LogStashLogger

Build Status Code Climate codecov.io Gem Version

LogStashLogger extends Ruby's Logger class to log directly to Logstash. It supports writing to various outputs in logstash JSON format. This is an improvement over writing to a file or syslog since Logstash can receive the structured data directly.

Features

  • Can write directly to a logstash listener over a UDP or TCP/SSL connection.
  • Can write to a file, Redis, Kafka, Kinesis, Firehose, a unix socket, syslog, stdout, or stderr.
  • Logger can take a string message, a hash, a LogStash::Event, an object, or a JSON string as input.
  • Events are automatically populated with message, timestamp, host, and severity.
  • Writes in logstash JSON format, but supports other formats as well.
  • Can write to multiple outputs.
  • Log messages are buffered and automatically re-sent if there is a connection problem.
  • Easily integrates with Rails via configuration.

Installation

Add this line to your application's Gemfile:

gem 'logstash-logger'

And then execute:

$ bundle

Or install it yourself as:

$ gem install logstash-logger

Usage Examples

require 'logstash-logger'

# Defaults to UDP on 0.0.0.0
logger = LogStashLogger.new(port: 5228)

# Specify host and type (UDP or TCP) explicitly
udp_logger = LogStashLogger.new(type: :udp, host: 'localhost', port: 5228)
tcp_logger = LogStashLogger.new(type: :tcp, host: 'localhost', port: 5229)

# Other types of loggers
file_logger = LogStashLogger.new(type: :file, path: 'log/development.log', sync: true)
unix_logger = LogStashLogger.new(type: :unix, path: '/tmp/sock')
syslog_logger = LogStashLogger.new(type: :syslog)
redis_logger = LogStashLogger.new(type: :redis)
kafka_logger = LogStashLogger.new(type: :kafka)
stdout_logger = LogStashLogger.new(type: :stdout)
stderr_logger = LogStashLogger.new(type: :stderr)
io_logger = LogStashLogger.new(type: :io, io: io)

# Use a different formatter
cee_logger = LogStashLogger.new(
  type: :tcp,
  host: 'logsene-receiver-syslog.sematext.com',
  port: 514,
  formatter: :cee_syslog
)

custom_formatted_logger = LogStashLogger.new(
  type: :redis,
  formatter: MyCustomFormatter
)

lambda_formatted_logger = LogStashLogger.new(
  type: :stdout,
  formatter: ->(severity, time, progname, msg) { "[#{progname}] #{msg}" }
)

ruby_default_formatter_logger = LogStashLogger.new(
  type: :file,
  path: 'log/development.log',
  formatter: ::Logger::Formatter
)

# Send messages to multiple outputs. Each output will have the same format.
# Syslog cannot be an output because it requires a separate logger.
multi_delegating_logger = LogStashLogger.new(
  type: :multi_delegator,
  outputs: [
    { type: :file, path: 'log/development.log' },
    { type: :udp, host: 'localhost', port: 5228 }
  ])

# Balance messages between several outputs.
# Works the same as multi delegator, but randomly chooses an output to send each message.
balancer_logger = LogStashLogger.new(
  type: :balancer,
  outputs: [
    { type: :udp, host: 'host1', port: 5228 },
    { type: :udp, host: 'host2', port: 5228 }
  ])

# Send messages to multiple loggers.
# Use this if you need to send different formats to different outputs.
# If you need to log to syslog, you must use this.
multi_logger = LogStashLogger.new(
  type: :multi_logger,
  outputs: [
    { type: :file, path: 'log/development.log', formatter: ::Logger::Formatter },
    { type: :tcp, host: 'localhost', port: 5228, formatter: :json }
  ])

# The following messages are written to UDP port 5228:

logger.info 'test'
# {"message":"test","@timestamp":"2014-05-22T09:37:19.204-07:00","@version":"1","severity":"INFO","host":"[hostname]"}

logger.error '{"message": "error"}'
# {"message":"error","@timestamp":"2014-05-22T10:10:55.877-07:00","@version":"1","severity":"ERROR","host":"[hostname]"}

logger.debug message: 'test', foo: 'bar'
# {"message":"test","foo":"bar","@timestamp":"2014-05-22T09:43:24.004-07:00","@version":"1","severity":"DEBUG","host":"[hostname]"}

logger.warn LogStash::Event.new(message: 'test', foo: 'bar')
# {"message":"test","foo":"bar","@timestamp":"2014-05-22T16:44:37.364Z","@version":"1","severity":"WARN","host":"[hostname]"}

# Tagged logging
logger.tagged('foo') { logger.fatal('bar') }
# {"message":"bar","@timestamp":"2014-05-26T20:35:14.685-07:00","@version":"1","severity":"FATAL","host":"[hostname]","tags":["foo"]}

URI Configuration

You can use a URI to configure your logstash logger instead of a hash. This is useful in environments such as Heroku where you may want to read configuration values from the environment. The URI scheme is type://host:port/path?key=value. Some sample URI configurations are given below.

udp://localhost:5228
tcp://localhost:5229
unix:///tmp/socket
file:///path/to/file
redis://localhost:6379
kafka://localhost:9092
stdout:/
stderr:/

Pass the URI into your logstash logger like so:

# Read the URI from an environment variable
logger = LogStashLogger.new(uri: ENV['LOGSTASH_URI'])

Logstash Listener Configuration

In order for logstash to correctly receive and parse the events, you will need to configure and run a listener that uses the json_lines codec. For example, to receive events over UDP on port 5228:

input {
  udp {
    host => "0.0.0.0"
    port => 5228
    codec => json_lines
  }
}

File and Redis inputs should use the json codec instead. For more information read the Logstash docs.

See the samples directory for more configuration samples.

SSL

If you are using TCP then there is the option of adding an SSL certificate to the options hash on initialize.

LogStashLogger.new(type: :tcp, port: 5228, ssl_certificate: "/path/to/certificate.crt")

The SSL certificate and key can be generated using

openssl req -x509 -batch -nodes -newkey rsa:2048 -keyout logstash.key -out logstash.crt

You can also enable SSL without a certificate:

LogStashLogger.new(type: :tcp, port: 5228, ssl_enable: true)

Specify an SSL context to have more control over the behavior. For example, set the verify mode:

ctx = OpenSSL::SSL::SSLContext.new
ctx.set_params(verify_mode: OpenSSL::SSL::VERIFY_NONE)
LogStashLogger.new(type: :tcp, port: 5228, ssl_context: ctx)

The following Logstash configuration is required for SSL:

input {
  tcp {
    host => "0.0.0.0"
    port => 5228
    codec => json_lines
    ssl_enable => true
    ssl_cert => "/path/to/certificate.crt"
    ssl_key => "/path/to/key.key"
  }
}

Hostname Verification

Hostname verification is enabled by default. Without further configuration, the hostname supplied to :host will be used to verify the server's certificate identity.

If you don't pass an :ssl_context or pass a falsey value to the :verify_hostname option, hostname verification will not occur.

Examples

Verify the hostname from the :host option

ctx = OpenSSL::SSL::SSLContext.new
ctx.cert = '/path/to/cert.pem'
ctx.verify_mode = OpenSSL::SSL::VERIFY_PEER

LogStashLogger.new \
  type: :tcp,
  host: 'logstash.example.com'
  port: 5228,
  ssl_context: ctx

Verify a hostname different from the :host option

LogStashLogger.new \
  type: :tcp,
  host: '1.2.3.4'
  port: 5228,
  ssl_context: ctx,
  verify_hostname: 'server.example.com'

Explicitly disable hostname verification

LogStashLogger.new \
  type: :tcp,
  host: '1.2.3.4'
  port: 5228,
  ssl_context: ctx,
  verify_hostname: false

Custom Log Fields

LogStashLogger by default will log a JSON object with the format below.

{
  "message":"Some Message",
  "@timestamp":"2015-01-29T10:43:32.196-05:00",
  "@version":"1",
  "severity":"INFO",
  "host":"hostname"
}

Some applications may need to attach additional metadata to each message. The LogStash::Event can be manipulated directly by specifying a customize_event block in the LogStashLogger configuration.

config = LogStashLogger.configure do |config|
  config.customize_event do |event|
    event["other_field"] = "some_other_value"
  end
end

This configuration would result in the following output.

{
    "message": "Some Message",
    "@timestamp": "2015-01-29T10:43:32.196-05:00",
    "@version": "1",
    "severity": "INFO",
    "host": "hostname",
    "other_field": "some_other_value"
}

This block has full access to the event, so you can remove fields, modify existing fields, etc. For example, to remove the default timestamp:

config = LogStashLogger.configure do |config|
  config.customize_event do |event|
    event.remove('@timestamp')
  end
end

You can also customize events on a per-logger basis by passing a callable object (lambda or proc) to the customize_event option when creating a logger:

LogStashLogger.new(customize_event: ->(event){ event['other_field'] = 'other_field' })

Buffering / Automatic Retries

For devices that establish a connection to a remote service, log messages are buffered internally and flushed in a background thread. If there is a connection problem, the messages are held in the buffer and automatically resent until it is successful. Outputs that support batch writing (Redis and Kafka) will write log messages in bulk from the buffer. This functionality is implemented using a fork of Stud::Buffer. You can configure its behavior by passing the following options to LogStashLogger:

  • :buffer_max_items - Max number of items to buffer before flushing. Defaults to 50.
  • :buffer_max_interval - Max number of seconds to wait between flushes. Defaults to 5.
  • :drop_messages_on_flush_error - Drop messages when there is a flush error. Defaults to false.
  • :drop_messages_on_full_buffer - Drop messages when the buffer is full. Defaults to true.
  • :sync - Flush buffer every time a message is received (blocking). Defaults to false.
  • :buffer_flush_at_exit - Flush messages when exiting the program. Defaults to true.
  • :buffer_logger - Logger to write buffer debug/error messages to. Defaults to none.

You can turn buffering off by setting sync = true.

Please be aware of the following caveats to this behavior:

  • It's possible for duplicate log messages to be sent when retrying. For outputs like Redis and Kafka that write in batches, the whole batch could get re-sent. If this is a problem, you can add a UUID field to each event to uniquely identify it. You can either do this in a customize_event block, or by using logstash's UUID filter.
  • It's still possible to lose log messages. Ruby won't detect a TCP/UDP connection problem immediately. In my testing, it took Ruby about 4 seconds to notice the receiving end was down and start raising exceptions. Since logstash listeners over TCP/UDP do not acknowledge received messages, it's not possible to know which log messages to re-send.
  • When sync is turned off, Ruby may buffer data internally before writing to the IO device. This is why you may not see messages written immediately to a UDP or TCP socket, even though LogStashLogger's buffer is periodically flushed.

Full Buffer

By default, messages are discarded when the buffer gets full. This can happen if the output source is down for too long or log messages are being received too quickly. If your application suddenly terminates (for example, by SIGKILL or a power outage), the whole buffer will be lost.

You can make message loss less likely by increasing buffer_max_items (so that more events can be held in the buffer), and decreasing buffer_max_interval (to wait less time between flushes). This will increase memory pressure on your application as log messages accumulate in the buffer, so make sure you have allocated enough memory to your process.

If you don't want to lose messages when the buffer gets full, you can set drop_messages_on_full_buffer = false. Note that if the buffer gets full, any incoming log message will block, which could be undesirable.

Sync Mode

All logger outputs support a sync setting. This is analogous to the "sync mode" setting on Ruby IO objects. When set to true, output is immediately flushed and is not buffered internally. Normally, for devices that connect to a remote service, buffering is a good thing because it improves performance and reduces the likelihood of errors affecting the program. For these devices, sync defaults to false, and it is recommended to leave the default value. You may want to turn sync mode on for testing, for example if you want to see log messages immediately after they are written.

It is recommended to turn sync mode on for file and Unix socket outputs. This ensures that log messages from different threads or proceses are written correctly on separate lines.

See #44 for more details.

Error Handling

If an exception occurs while writing a message to the device, the exception is logged using an internal logger. By default, this logs to $stderr. You can change the error logger by setting LogStashLogger.configuration.default_error_logger, or by passsing your own logger object in the :error_logger configuration key when instantiating a LogStashLogger.

Logger Silencing

LogStashLogger provides support for Rails-style logger silencing. The implementation was extracted from Rails, but has no dependencies, so it can be used outside of a Rails app. The interface is the same as in Rails:

logger.silence(temporary_level) do
  ...
end

Custom Logger Class

By default, LogStashLogger creates a logger that extends Ruby's built in Logger class. If you require a different logger implementation, you can use a different class by passing in the class with the logger_class option.

Note that for syslog, the Syslog::Logger class is required and cannot be changed.

Rails Integration

Supports Rails 4.2 and 5.x.

By default, every Rails log message will be written to logstash in LogStash::Event JSON format.

For minimal, more-structured logstash events, try one of the following gems:

Currently these gems output a JSON string, which LogStashLogger then parses. Future versions of these gems could potentially have deeper integration with LogStashLogger (e.g. by directly writing LogStash::Event objects).

Rails Configuration

Add the following to your config/environments/production.rb:

Common Options

# Optional, Rails sets the default to :info
config.log_level = :debug

# Optional, Rails 4 defaults to true in development and false in production
config.autoflush_log = true

# Optional, use a URI to configure. Useful on Heroku
config.logstash.uri = ENV['LOGSTASH_URI']

# Optional. Defaults to :json_lines. If there are multiple outputs,
# they will all share the same formatter.
config.logstash.formatter = :json_lines

# Optional, the logger to log writing errors to. Defaults to logging to $stderr
config.logstash.error_logger = Logger.new($stderr)

# Optional, max number of items to buffer before flushing. Defaults to 50
config.logstash.buffer_max_items = 50

# Optional, max number of seconds to wait between flushes. Defaults to 5
config.logstash.buffer_max_interval = 5

# Optional, drop message when a connection error occurs. Defaults to false
config.logstash.drop_messages_on_flush_error = false

# Optional, drop messages when the buffer is full. Defaults to true
config.logstash.drop_messages_on_full_buffer = true

UDP

# Optional, defaults to '0.0.0.0'
config.logstash.host = 'localhost'

# Optional, defaults to :udp.
config.logstash.type = :udp

# Required, the port to connect to
config.logstash.port = 5228

TCP

# Optional, defaults to '0.0.0.0'
config.logstash.host = 'localhost'

# Required, the port to connect to
config.logstash.port = 5228

# Required
config.logstash.type = :tcp

# Optional, enables SSL
config.logstash.ssl_enable = true

Unix Socket

# Required
config.logstash.type = :unix

# Required
config.logstash.path = '/tmp/sock'

Syslog

If you're on Ruby 1.9, add Syslog::Logger v2 to your Gemfile:

gem 'SyslogLogger', '2.0'

If you're on Ruby 2+, Syslog::Logger is already built into the standard library.

# Required
config.logstash.type = :syslog

# Optional. Defaults to 'ruby'
config.logstash.program_name = 'MyApp'

# Optional default facility level. Only works in Ruby 2+
config.logstash.facility = Syslog::LOG_LOCAL0

Redis

Add the redis gem to your Gemfile:

gem 'redis'
# Required
config.logstash.type = :redis

# Optional, will default to the 'logstash' list
config.logstash.list = 'logstash'

# All other options are passed in to the Redis client
# Supported options include host, port, path, password, url
# Example:

# Optional, Redis will default to localhost
config.logstash.host = 'localhost'

# Optional, Redis will default to port 6379
config.logstash.port = 6379

Kafka

Add the poseidon gem to your Gemfile:

gem 'poseidon'
# Required
config.logstash.type = :kafka

# Optional, will default to the 'logstash' topic
config.logstash.path = 'logstash'

# Optional, will default to the 'logstash-logger' producer
config.logstash.producer = 'logstash-logger'

# Optional, will default to localhost:9092 host/port
config.logstash.hosts = ['localhost:9092']

# Optional, will default to 1s backoff
config.logstash.backoff = 1

Kinesis

Add the aws-sdk gem to your Gemfile:

# aws-sdk >= 3.0
gem 'aws-sdk-kinesis'

# aws-sdk < 3.0
gem 'aws-sdk'
# Required
config.logstash.type = :kinesis

# Optional, will default to the 'logstash' stream
config.logstash.stream = 'my-stream-name'

# Optional, will default to 'us-east-1'
config.logstash.aws_region = 'us-west-2'

# Optional, will default to the AWS_ACCESS_KEY_ID environment variable
config.logstash.aws_access_key_id = 'ASKASKHLD12341'

# Optional, will default to the AWS_SECRET_ACCESS_KEY environment variable
config.logstash.aws_secret_access_key = 'ASKASKHLD1234123412341234'

Firehose

Add the aws-sdk gem to your Gemfile:

# aws-sdk >= 3.0
gem 'aws-sdk-firehose'

# aws-sdk < 3.0
gem 'aws-sdk'
# Required
config.logstash.type = :firehose

# Optional, will default to the 'logstash' delivery stream
config.logstash.stream = 'my-stream-name'

# Optional, will default to AWS default region config chain
config.logstash.aws_region = 'us-west-2'

# Optional, will default to AWS default credential provider chain
config.logstash.aws_access_key_id = 'ASKASKHLD12341'

# Optional, will default to AWS default credential provider chain
config.logstash.aws_secret_access_key = 'ASKASKHLD1234123412341234'

File

# Required
config.logstash.type = :file

# Optional, defaults to Rails log path
config.logstash.path = 'log/production.log'

IO

# Required
config.logstash.type = :io

# Required
config.logstash.io = io

Multi Delegator

# Required
config.logstash.type = :multi_delegator

# Required
config.logstash.outputs = [
  {
    type: :file,
    path: 'log/production.log'
  },
  {
    type: :udp,
    port: 5228,
    host: 'localhost'
  }
]

Multi Logger

# Required
config.logstash.type = :multi_logger

# Required. Each logger may have its own formatter.
config.logstash.outputs = [
  {
    type: :file,
    path: 'log/production.log',
    formatter: ::Logger::Formatter
  },
  {
    type: :udp,
    port: 5228,
    host: 'localhost'
  }
]

Logging HTTP request data

In web applications, you can log data from HTTP requests (such as headers) using the RequestStore middleware. The following example assumes Rails.

# in Gemfile
gem 'request_store'
# in application.rb
LogStashLogger.configure do |config|
  config.customize_event do |event|
    event["session_id"] = RequestStore.store[:load_balancer_session_id]
  end
end
# in app/controllers/application_controller.rb
before_filter :track_load_balancer_session_id

def track_load_balancer_session_id
  RequestStore.store[:load_balancer_session_id] = request.headers["X-LOADBALANCER-SESSIONID"]
end

Cleaning up resources when forking

If your application forks (as is common with many web servers) you will need to manage cleaning up resources on your LogStashLogger instances. The instance method #reset is available for this purpose. Here is sample configuration for several common web servers used with Rails:

Passenger:

::PhusionPassenger.on_event(:starting_worker_process) do |forked|
  Rails.logger.reset
end

Puma:

# In config/puma.rb
on_worker_boot do
  Rails.logger.reset
end

Unicorn

# In config/unicorn.rb
after_fork do |server, worker|
  Rails.logger.reset
end

Ruby Compatibility

Verified to work with:

  • MRI Ruby 2.2 - 2.5
  • JRuby 9.x
  • Rubinius

Ruby versions < 2.2 are EOL'ed and no longer supported.

What type of logger should I use?

It depends on your specific needs, but most applications should use the default (UDP). Here are the advantages and disadvantages of each type:

  • UDP is faster than TCP because it's asynchronous (fire-and-forget). However, this means that log messages could get dropped. This is okay for many applications.
  • TCP verifies that every message has been received via two-way communication. It also supports SSL for secure transmission of log messages over a network. This could slow your app down to a crawl if the TCP listener is under heavy load.
  • A file is simple to use, but you will have to worry about log rotation and running out of disk space.
  • Writing to a Unix socket is faster than writing to a TCP or UDP port, but only works locally.
  • Writing to Redis is good for distributed setups that generate tons of logs. However, you will have another moving part and have to worry about Redis running out of memory.
  • Writing to stdout is only recommended for debugging purposes.

For a more detailed discussion of UDP vs TCP, I recommend reading this article: UDP vs. TCP

Troubleshooting

Logstash never receives any logs

If you are using a device backed by a Ruby IO object (such as a file, UDP socket, or TCP socket), please be aware that Ruby keeps its own internal buffer. Despite the fact that LogStashLogger buffers messages and flushes them periodically, the data written to the IO object can be buffered by Ruby internally indefinitely, and may not even write until the program terminates. If this bothers you or you need to see log messages immediately, your only recourse is to set the sync: true option.

JSON::GeneratorError

Your application is probably attempting to log data that is not encoded in a valid way. When this happens, Ruby's standard JSON library will raise an exception. You may be able to overcome this by swapping out a different JSON encoder such as Oj. Use the oj_mimic_json gem to use Oj for JSON generation.

No logs getting sent on Heroku

Heroku recommends installing the rails_12factor so that logs get sent to STDOUT. Unfortunately, this overrides LogStashLogger, preventing logs from being sent to their configured destination. The solution is to remove rails_12factor from your Gemfile.

Logging eventually stops in production

This is most likely not a problem with LogStashLogger, but rather a different gem changing the log level of Rails.logger. This is especially likely if you're using a threaded server such as Puma, since gems often change the log level of Rails.logger in a non thread-safe way. See #17 for more information.

Sometimes two lines of JSON log messages get sent as one message

If you're using UDP output and writing to a logstash listener, you are most likely encountering a bug in the UDP implementation of the logstash listener. There is no known fix at this time. See #43 for more information.

Errno::EMSGSIZE - Message too long

A known drawback of using TCP or UDP is the 65535 byte limit on total message size. To workaround this issue, you will have to truncate the message by setting the max message size:

LogStashLogger.configure do |config|
  config.max_message_size = 2000
end

This will truncate only the message field of the LogStash Event. So make sure you set the max message size significantly less than 65535 bytes to make room for other fields.

Breaking changes

Version 0.25+

Rails 3.2, MRI Ruby < 2.2, and JRuby 1.7 are no longer supported, since they have been EOL'ed. If you are on an older version of Ruby, you will need to use 0.24 or below.

Version 0.5+

  • The source event key has been replaced with host to better match the latest logstash.
  • The (host, port, type) constructor has been deprecated in favor of an options hash constructor.

Version 0.4+

LogStash::Event uses the v1 format starting version 1.2+. If you're using the v1, you'll need to install LogStashLogger version 0.4+. This is not backwards compatible with the old LogStash::Event v1.1.5, which uses the v0 format.

Version 0.3+

Earlier versions of this gem (<= 0.2.1) only implemented a TCP connection. Newer versions (>= 0.3) also implement UDP, and use that as the new default. Please be aware if you are using the default constructor and still require TCP, you should add an additional argument:

# Now defaults to UDP instead of TCP
logger = LogStashLogger.new('localhost', 5228)
# Explicitly specify TCP instead of UDP
logger = LogStashLogger.new('localhost', 5228, :tcp)

Contributors

Contributing

  1. Fork it
  2. Create your feature branch (git checkout -b my-new-feature)
  3. Commit your changes (git commit -am 'Add some feature')
  4. Push to the branch (git push origin my-new-feature)
  5. Create new Pull Request