peastash

Peastash

Description

Peastash helps you instrument your Ruby code with the ELK stack. It provides useful methods to easily write logs in a format that Logstash can understand, which will turn in (hopefully) beautiful dashboards using Elasticsearch and Kibana.

The design philosophy behind Peastash is:

• Be as unobstrusive as possible. Your code shouldn't be riddled with references to Peastash.
• Keep the code-base small and well-tested
• Provide sane default that will work for most people out of the box

Any ruby application can benefit from having clear metrics, and now you have one less excuse to start instrumenting your code today.

Usage

Installing

Add this line to your application's Gemfile:

gem 'peastash'

Getting started

Peastash provides a simple interface to aggregate data to be sent to Logstash. The boundaries of a log entry are defined by the Peastash.with_instance.log block. When the code inside the block is done running, a log entry will be created and written. The most basic usage would look like:

Peastash.with_instance.log do |instance|
  instance.store[:foo] = 'bar'
end
# This will produce a Logstash log entry looking like this:
# {"@source":"peastash","@fields":{"foo":"bar"},"@tags":[],"@timestamp":"2014-05-27T15:18:29.824Z","@version":"1"}

Configuration

Peastash ships with sane defaults, but can be easily configured:

Peastash.with_instance.configure!({
  source: 'peastash', # This value will be used for the @source field of the logstash event
  # Any tag you wish to find in Logstash later (tags passed as argument to #log are added).
  tags: [], # Defaults to [].
  # Dumps to STDOUT by default. Any object answering to the dump method
  output: Peastash::Outputs::IO.new('/tmp/peastash.log'),
  store_name: 'peastash', # The key used to reference the store in the Thread.current
  dump_if_empty: true # Whether or not Peastash should produce a log entry when the store is empty
})

Safety

Peastash is safe by default. This means that if anything raises, Peastash will keep on going with the program's execution flow to be as unobtrusive as possible. If this behaviour is unsuitable for your needs (for example if you want to test the new Peastash output you've just developped), simply call:

Peastash.unsafe!

Please note that this is a global setting and cannot be set per-instance.

Outputs

Peastash ships with a single output: Peastash::Outputs::IO. It can be initialized either by passing it a path or an IO object. If the argument is a path to a non-writeable file, Peastash will attempt to delete the file to re-create it with the proper permissions.

Peastash can easily be extended to output to any target. Simply configure Peastash's output with an object that responds to the #dump method. This method will be called at the end of the #log block, with 1 argument: a LogStash::Event object that you will probably need to serialize to json.

In case the directory doesn't exist, peastash won't crash, a tempfile will be created in /tmp (or $TMPDIR) and the data will be stored in this file. The full path of this tempfile will be printed to STDERR so it can easily be retrieved.

What if I want to use it in my rack app?

There's a middleware for that! Simply add the following somewhere:

require 'peastash/middleware'
use Peastash::Middleware

By default, the middleware only adds the duration, status and hostname (machine name) fields to the log entry. In addition to using Peastash.with_instance.store to add information, you can pass one or two block arguments to the use dsl, that will be called with the request env and the Rack response in parameter, in the context of the Middleware's instance. The first block will be called before the request (with a response = nil), while the second one will be called after, with the actual response. For example:

require 'peastash/middleware'
before_block = ->(env, response) { Peastash.with_instance.store[:path] = Rack::Request.new(env).path }
after_block = ->(env, response) { Peastash.with_instance.store[:headers] = response[1] }
use Peastash::Middleware, before_block, after_block
# Will add 'path' and headers to the list of fields.

Any instance variable you set in before_block will be available in after_block, but those instances variables WILL NOT be reset at the end of the request.

But I want to use it in Rails!

It's easy! Simply add Peastash to your Gemfile, and add the following to the configure block of either config/environment.rb or config/<RAILS_ENV>.rb:

config.peastash.enabled = true
# You can also configure Peastash from here, for example:
config.peastash.output = Peastash::Outputs::IO.new(File.join(Rails.root, 'log', "logstash_#{Rails.env}.log"))
config.peastash.source = Rails.application.class.parent_name
config.peastash.before_block = ->(env, response) { Peastash.with_instance.store[:path] = Rack::Request.new(env).path }
config.peastash.after_block = ->(env, response) { Peastash.with_instance.store[:puma_wait] = env['puma.request_body_wait'] }

By default, Peastash's Rails integration will log the same parameters as the Middleware version, plus the fields in the payload of the process_action.action_controller notification (except the params).

All the options for Peastash can be set using the config.peastash configuration object.

Logging request parameters

To enable parameter logging, you must add the following to your configuration:

config.peastash.log_parameters = true

Be careful, as this can significantly increase the size of the log entries, as well as causing issues if other Logstash entries have the same field with a different data type.

Listening to ActiveSupport::Notifications

Additionaly, you can use Peastash to aggregate data from any ActiveSupport::Notifications:

# In config/initializers/peastash.rb
if defined?(Peastash.with_instance.watch)
  Peastash.with_instance.watch('request.rsolr', event_group: 'solr') do |name, start, finish, id, payload, event_store|
    event_store[:queries] = event_store[:queries].to_i.succ
    event_store[:duration] = event_store[:duration].to_f + ((finish - start) * 1000)
  end
end
# This will add something like the following to the log entry fields
# {'request.rsolr': {'queries': 1, 'duration': 42}}

The store exposed to the blocked passed to watch is thread-safe, and is reset after each request. By default, the store is only shared between occurences of the same event. You can easily share the same store between different types of notifications by assigning them to the same event group:

Peastash.with_instance.watch('foo.notification', event_group: 'notification') do |*args, store|
  # Shared store with 'bar.notification'
end

Peastash.with_instance.watch('bar.notification', event_group: 'notification') do |*args, store|
  # Shared store with 'foo.notification'
end

Playing with multiple instances

Calling Peastash.with_instance is the same as calling Peastash.with_instance(:global). You can use different instances to log different things. Each instance as its own:

• Configuration (inherited from the global instance)
• Store
• Tags

Using several instances, you can nest log blocks without sharing the store, or simply use them for different purposes. For example, you can log your Rails query with the global instance, and your asynchronous jobs with a worker instance, and have those instances output to different files.

Playing with tags

There are three ways to tag your log entries in Peastash.with_instance.

• The first one is through configuration (see above)
• The second one is by passing tags to the Peastash.with_instance.log method
• The last one is to call the Peastash.with_instance.tags method from within a Peastash.with_instance.log block. Tags defined with this method are not persistent and will disappear at the next call to Peastash.with_instance.log

For example:

Peastash.with_instance.configure!(tags: 'foo')
Peastash.with_instance.log(['bar']) { Peastash.with_instance.tags << 'baz' }
# Tags are ['foo', 'bar', 'baz']
Peastash.with_instance.log(['bar']) { Peastash.with_instance.tags << 'qux' }
# Tags are ['foo', 'bar', 'qux']

Contributing

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

Don't forget to run the tests with rake.