tty logo

TTY::Logger Gitter

Gem Version Actions CI Build status Code Climate Coverage Status Inline docs

A readable, structured and beautiful logging for the terminal

TTY::Logger provides independent logging component for TTY toolkit.


  • Intuitive console output for an increased readability
  • Ability to stream data to any IO object
  • Supports structured data logging
  • Filters sensitive data
  • Allows to define custom log types
  • Formats and truncates messages to avoid clogging logging output
  • Customizable styling of labels and symbols for console output
  • Includes metadata information: time, location, scope
  • Handles multiple logging outputs


Add this line to your application's Gemfile:

gem "tty-logger"

And then execute:

$ bundle

Or install it yourself as:

$ gem install tty-logger


1. Usage

Create logger:

logger =

And log information using any of the logger built-in types: "Deployed successfully" "Deployed", "successfully" { "Dynamically generated info" }

Include structured data:

logger.success "Deployed successfully", myapp: "myapp", env: "prod"
# =>
# ✔ success Deployed successfully     app=myapp env=prod

Add metadata information:

logger = do |config|
  config. = [:date, :time]
end "Deployed successfully", myapp: "myapp", env: "prod"
# =>
# [2019-07-17] [23:21:55.287] › ℹ info    Info about the deploy     app=myapp env=prod

Or change structured data formatting display to JSON:

logger = do |config|
  config.formatter = :json
end "Deployed successfully"
# =>
# [2019-07-17] [23:21:55.287] › ℹ info    Info about the deploy     {"app":"myapp","env":"prod"}

2. Synopsis

2.1 Logging

There are many logger types to choose from:

  • debug - logs message at :debug level
  • info - logs message at :info level
  • success - logs message at :info level
  • wait - logs message at :info level
  • warn - logs message at :warn level
  • error - logs message at :error level
  • fatal - logs message at :fatal level

To log a message, simply choose one of the above types and pass in the actual message. For example, to log successfully deployment at info level do:

logger.success "Deployed successfully"
# =>
# ✔ success Deployed successfully

Or pass in multiple messages:

logger.success "Deployed", "successfully"
# =>
# ✔ success Deployed successfully

You can delay message evaluation by passing it inside a block:

logger.success { "Dynamically generated info" }
# =>
# ✔ success Dynamically generated info

Similar to regular logging, you cal split your message into chunks inside a block:

logger.success { ["Dynamically", "generated", "info"] }
# =>
# ✔ success Dynamically generated info

The above comes handy when paired with structured data.

2.1.1 Exceptions

You can also report on exceptions.

For example, let's say you caught an exception about incorrect data format and use fatal level to log it:

  raise ArgumentError, "Wrong data"
rescue => ex
  logger.fatal("Error:", ex)

This will result in a message followed by a full backtrace:

# =>
# ! fatal   Error: Wrong data
#    tty-logger/spec/unit/exception_spec.rb:12:in `block (2 levels) in <top (required)>'
#    rspec-core-3.8.2/lib/rspec/core/example.rb:257:in `instance_exec'
#    rspec-core-3.8.2/lib/rspec/core/example.rb:257:in `block in run'

2.1.2 Types

You can define custom log types via the types configuration option:

For example, if you want to add thanks and done log types, you need to provide their names along with logging levels. You can further customise the :console output with your desired styling:

logger = do |config|
  config.types = {
    thanks: {level: :info},
    done: {level: :info}
  config.handlers = [
    [:console, {
      styles: {
        thanks: {
          symbol: "❤️ ",
          label: "thanks",
          color: :magenta,
          levelpad: 0
        done: {
          symbol: "!!",
          label: "done",
          color: :green,
          levelpad: 2

Once defined, you can call new log types:

logger.thanks("Great work!")
logger.done("Work done!")
# =>
# ❤️  thanks Great work!
# !! done   Work done!

2.2 Levels

The supported levels, ordered by precedence, are:

  • :debug - for debug-related messages
  • :info - for information of any kind
  • :warn - for warnings
  • :error - for errors
  • :fatal - for fatal conditions

So the order is: :debug < :info < :warn < :error < :fatal

For example, :info takes precedence over :debug. If your log level is set to :info, :info, :warn, :error and :fatal will be printed to the console. If your log level is set to :warn, only :warn, :error and :fatal will be printed.

You can set level using the level configuration option. The value can be a symbol, a string or level constant. For example, :info, INFO or TTY::Logger::INFO_LEVEL will qualify as valid level value. do |config|
  config.level = :info # or "INFO" or TTY::Logger::INFO_LEVEL

Or you can specific level for each log events handler.

For example, to log messages above :info level to a stream and only :error level events to the console do:

logger = do |config|
  config.handlers = [
    [:console, level: :error],
    [:stream, level: :info]

You can also change the output streams for each handler.

2.2.1 Scoped Level

You can temporarily change level, raise it or lower it by using the log_at call. By default messages are logged at :info level, but you can change this for the duration of a block:

logger ="not logged")

logger.log_at :debug do
# =>
# • debug   logged

Or elevate a level to an error with a constant ERROR_LEVEL:

logger.log_at TTY::Logger::ERROR_LEVEL do
  logger.debug("not logged")
# =>
# ⨯ error   logged

2.3 Structured data

To add global data available for all logger calls:

logger = {app: "myapp", env: "prod"})"Deploying...")
# =>
# ℹ info    Deploying...              app=myapp env=prod

To only add data for a single log event:

logger =
logger.wait "Ready to deploy", app: "myapp", env: "prod"
# =>
# … waiting Ready to deploy           app=myapp env=prod

You can delay data evaluation until it's evaluated by passing it inside a block:

logger.wait { ["Ready to deploy", {app: "myapp", env: "prod"}] }
# =>
# … waiting Ready to deploy           app=myapp env=prod

2.4 Configuration

All the configuration options can be changed globally via configure or per logger instance.

  • :filters - the storage of placeholders to filter sensitive data out from the logs. Defaults to {}.
  • :formatter - the formatter used to display structured data. Defaults to :text. See Formatters for more details.
  • :handlers - the handlers used to log messages. Defaults to [:console]. See Handlers for more details.
  • :level - the logging level. Any message logged below this level will be simply ignored. Each handler may have it's own default level. Defaults to :info
  • :max_bytes - the maximum message size to be logged in bytes. Defaults to 8192 bytes. The truncated message will have ... at the end.
  • :max_depth - the maximum depth for nested structured data. Defaults to 3.
  • :metadata - the meta info to display before the message, can be :pid, :date, :time or :file. Defaults to empty array [], no metadata. Setting this to :all will print all the metadata.
  • :types - the new custom log types. Defaults to {}.
  • :date_format - uses strftime format to display dates. Defaults to "%F".
  • :time_format - uses strftime format to display times. Defaults to "%T.%3N".

For example, to configure :max_bytes, :level and :metadata for all logger instances do:

TTY::Logger.configure do |config|
  config.max_bytes = 2**10
  config.level = :error
  config. = [:time, :date]

Or if you wish to setup configuration per logger instance use block:

logger = do |config|
  config.max_bytes = 2**20
  config. = [:all]

You can also change the logger's configuration at runtime:

logger.configure do |config|
  config.level = :debug

2.4.1 Metadata

The :metdata configuration option can include the following symbols:

  • :pid - the log event process identifier
  • :date - the log event date
  • :time - the log event time
  • :file - the file with a line number the log event is triggered from

2.4.2 Filters

You can filter sensitive data out of log output with filters configuration option. The filters can be further configured to remove info from log message with message or structured data with data. Both methods, as a value accept a list of sensitive items to search for.

If you want to filter sensitive information from log messages use message:

logger = output) do |config|
  config.filters.message = %w[secret password]

Which by default will replace each matching string with [FILTERED] placeholder:"Super secret info with password")
# =>
# ℹ info    Super [FILTERED] info with [FILTERED]

You can also replace each data item with a custom placeholder. To do so use a :mask keyword with a replacement placeholder.

For example, to replace "secret" content with placeholder "<SECRET>" do:

logger = do |config|
  config.filters.message = %w[secret]
  config.filters.mask = "<SECRET>"

When logged, it will produce:"Super secret info")
# =>
# ℹ info    Super <SECRET> info

To filter out sensitive information out of structured data use data method. By default any value matching a parameter name will be filtered regardless of the level of nesting. If you wish to filter only a specific deeply nested key use a dot notation like params.card.password to only filter {params: {card: {password: "Secret123"}}}.

For example to filter out a :password from data do:

logger = do |config| = i[password]

This will filter out any key matching password:"Secret info", password: "Secret123", email: "")
# =>
# ℹ info    Secret info     password="[FILTERED]" email="[email protected]"

But also any nested data item:"Secret info", params: {password: "Secret123", email: ""})
# =>
# ℹ info    Secret info     params={password="[FILTERED]" email="[email protected]"}

You're not limited to using only direct string comparison. You can also match based on regular expressions. For example, to match keys starting with ba we can add a following filter:

logger = do |config| = [/ba/]

Then appropriate values will be masked:"Filtering data", {"foo" => {"bar" => "val", "baz" => "val"}})
# =>
# ℹ info    Filtering data            foo={bar="[FILTERED]" baz="[FILTERED]"}

You can mix and match. To filter keys based on pattern inside a deeply nested hash use dot notation with regular expression. For example, to find keys for the :foo parent key that starts with :b character, we could do:

logger = do |config| = [/^foo\.b/]

Then only keys under the :foo key will be filtered:"Filtering data", {"foo" => {"bar" => "val"}, "baz" => {"bar" => val"}})
# =>
# ℹ info    Filtering data            foo={bar="[FILTERED]"} baz={bar=val}

2.5 Cloning

You can create a copy of a logger with the current configuration using the copy method.

For example, given the following logger with :app and :env data:

logger = {app: "parent", env: "prod"})

We can create a copy with a custom configuration that changes filtered message content and :app data:

child_logger = logger.copy(app: "child") do |config|
  config.filters = ["logging"]
end"Parent logging")
child_logger.warn("Child logging")
# =>
# ℹ info    Parent logging            app=parent env=prod
# ⚠ warning Child [FILTERED]          app=child env=prod

2.6 Handlers

TTY::Logger supports many ways to handle log messages.

The available handlers by default are:

  • :console - log messages to the console, enabled by default
  • :null - discards any log messages
  • :stream - log messages to an IO stream, a file, a socket or a console.

You can also implement your own custom handler.

The handlers can be configured via global or instance configuration with handlers. The handler can be a name or a class name: do |config|
  config.handlers = [:console]

Or using class name: do |config|
  config.handlers = [TTY::Logger::Handlers::Console]

Handlers can also be added/removed dynamically through add_handler or remove_handler.

logger =

2.6.1 Console Handler

The console handler prints log messages to the console. It supports the following options:

  • :styles - a hash of styling options.
  • :formatter - the formatter for log messages. Defaults to :text.
  • :output - the device to log error messages to. Defaults to $stderr.
  • :message_format - uses sprintf format to display messages. Defaults to "%-25s".
  • :enable_color - when true forces colored output, when false disables colored output. Defaults to nil which performs automatic terminal color support detection.

The supported options in the :styles are:

  • :label - the name for the log message.
  • :symbol - the graphics to display before the log message label.
  • :color - the color for the log message.
  • :levelpad - the extra amount of padding used to display log label.

See the TTY::Logger::Handlers::Console for full list of styles.

Console handler has many default styles such as success and error:

logger =
logger.success("Default success")
logger.error("Default error")
# =>
# ✔ success Default success
# ⨯ error   Default error

You can change the default styling with a tuple of handler name and options hash.

In our example, we want to change the styling of success and error:

new_styles = {
  styles: {
    success: {
      symbol: "+",
      label: "Ohh yes"
    error: {
      symbol: "!",
      label: "Dooh",
      levelpad: 3 # the amount of extra padding to align level names in a column

And then use the new_styles when providing handlers configuration:

styled_logger = do |config|
  config.handlers = [[:console, new_styles]]

styled_logger.success("Custom success")
styled_logger.error("Custom error")
# =>
# + Ohh yes Custom success
# ! Dooh    Custom error

To increase message padding to a percentage of terminal width (depends on tty-screen): do |config|
  padding = (TTY::Screen.columns * 0.4).to_i
  config.handlers = [[:console, { message_format: "%-#{padding}s" }]]

2.6.2 Stream handler

To send log event data outside of console to another service or IO stream, you can use :stream handler.

logger = output) do |config|
  config.handlers = [:stream]
  config. = [:all]

By default, the output will be a plain text streamed to console. The text contains key and value pairs of all the metadata and the message of the log event."Info about the deploy", app:"myap", env:"prod")
# =>
# pid=18315 date="2019-07-21" time="15:42:12.463" path="examples/stream.rb:17:in`<main>`"
# level=info message="Info about the deploy" app=myapp env=prod

You can change stream formatter for ease of working with external services such as Logstash. For example, to use :stream handler with :json format do:

logger = output) do |config|
  config.handlers = [[:stream, formatter: :json]]
  config. = [:all]

This will output JSON formatted text streamed to console."Info about the deploy", app="myap", env="prod")
# =>
# {"pid":18513,"date":"2019-07-21","time":"15:54:09.924","path":"examples/stream.rb:17:in`<main>`",
# "level":"info","message":"Info about the deploy","app":"myapp","env":"prod"}

2.6.3 Custom Handler

You can create your own log event handler if the default ones don't match your needs.

The design of your handler should include two calls:

  • initialize - where all dependencies get injected
  • call - where the log event is handled

We start with the implementation of the initialize method. This method by default is injected with :config key that includes all global configuration options. The :output key for displaying log message in the console and :formatter.

In our case we also add custom :label:

class MyHandler
  def initialize(output: nil, config: nil, formatter: nil, label: nil)
    @label = label
    @output = output

Next is the call method that accepts the log event.

The event has the following attributes:

  • message - the array of message parts to be printed
  • fields - the structured data supplied with the event
  • metadata - the additional info about the event. See metadata section for details.

We add implementation of call:

class MyHandler
  def initialize(output: nil, config: nil, label: nil)
    @label = label
    @output = output

  def call(event)
    @output.puts "(#{@label}) #{event.message.join}"

Once you have your custom handler, you need to register it with the logger. You can do so using the handlers configuration option:

logger = do |config|
  config.handlers = [[MyHandler, label: "myhandler"]]

Or add your handler dynamically after logger initialization:

logger =
logger.add_handler [MyHandler, label: "myhandler"]

2.6.4 Multiple Handlers

You can define as many handlers as you need. For example, you may log messages both to console and stream:

logger = do |config|
  config.handlers = [:console, :stream]

Each handler can have its own configuration. For example, you can register :console handler to log messages above error level and :stream that logs any message with info or higher level:

logger = do |config|
  config.handlers = [
    [:console, level: :error],
    [:stream, level: :info]

2.7 Formatters

The available formatters are:

  • :json
  • :text

You can configure format for all the handlers: do |config|
  config.formatter = :json

Or specify a different formatter for each handler. For example, let's say you want to log to console twice, once with default formatter and once with :json formatter: do |config|
  config.handlers = [:console, [:console, formatter: :json]]

2.8 Output Streams

By default all log events are output to stderr. You can change this using configuration output option. Any IO-like stream such as file, socket or console can be used. For example, to log all messages to a file do:

logger = do |config|
  config.output ="errors.log", "a")

You can also specify multiple streams that all log messages will be sent to:

logger = do |config|
  config.output = [$stderr,"errors.log", "a")]

Conversely, you can specify different output for each of the handlers used. For example, you can output all messages above info level to a file with a stream handler and only show error messages in the console with a nicely formatted output.

logger = do |config|
  config.handlers = [
    [:console, output: $stderr, level: :error],
    [:stream, output:"errors.log", "a"), level: :info)]

3. Community Extensions

3.1 Sentry Handler

tty-logger-raven provides an extension for


After checking out the repo, run bin/setup to install dependencies. Then, run rake spec to run the tests. You can also run bin/console for an interactive prompt that will allow you to experiment.

To install this gem onto your local machine, run bundle exec rake install. To release a new version, update the version number in version.rb, and then run bundle exec rake release, which will create a git tag for the version, push git commits and tags, and push the .gem file to


Bug reports and pull requests are welcome on GitHub at This project is intended to be a safe, welcoming space for collaboration, and contributors are expected to adhere to the Contributor Covenant code of conduct.


The gem is available as open source under the terms of the MIT License.

Code of Conduct

Everyone interacting in the TTY::Logger project’s codebases, issue trackers, chat rooms and mailing lists is expected to follow the code of conduct.

Copyright (c) 2019 Piotr Murach. See LICENSE for further details.