Skip to content

Latest commit

 

History

History
342 lines (261 loc) · 11.5 KB

README.md

File metadata and controls

342 lines (261 loc) · 11.5 KB

Listerine

Listerine is a simple functional monitoring framework that enables you to quickly create customizable functional monitors with full alerting capability.

Listerine is not a process monitoring framework like God but instead intended for functional test application monitoring.

This project was originally created as a self-service for Appboy as a replacement for more expensive services. Read the blog post introducing it.

Listerine enables you to monitor all levels of your web applications and services. Some common examples include:

  • Ensure that your caching layer is functioning properly
  • Make sure that you have X available Resque workers
  • Make sure that your Resque Scheduler agent is online
  • Check that your hosted database is online (e.g., if you use MongoHQ or RedisToGo)
  • POST to your API and verify return values

Installation

gem install listerine

Overview

Listerine allows you to define simple script monitors that contain an assertion. When the assertion is true, the monitor has succeeded. When the assertion evaluates to false the monitor is marked as failed, sends a notification, and can run optional code on failure.

All monitors must contain both name and assert blocks. Unhandled exceptions are caught and treated as failures, with the exception text and backtrace included in the notification. Here's an example:

require 'mysql2'

# Global configuration settings
Listerine::Monitor.configure do
  # Configure the email address from which the alerts will be sent
  from "[email protected]"

  # Notify [email protected] on all failures
  notify "[email protected]"
end

# Define the monitor
Listerine::Monitor.new do
  name "Database online"
  description "This monitor ensures that the database is online."
  assert do
    # Connect to MySQL. If the host is down, this will raise a Mysql2::Error exception, which will automatically
    # be caught by the Listerine::Runner handler, treated as a failure, and the exception text and backtrace will
    # be sent in the notification.
    Mysql2::Client.new(:host => "host", :username => "test_user")
    true
  end
end

# Run all the monitors -- declare all monitors before this line
Listerine::Runner.instance.run

If you were to run this file, the output would look like:

* Database online	PASS

To run this file regularly, schedule a cron job such as the below to run monitors every 2 minutes.

*/2 * * * * /path/to/monitor/file.rb

Multiple environments

The same monitor can be run in multiple environments. To do so, specify a list of environments when defining the monitor. In your assert block, you will have access to a method current_environment which indicates the environment in which the monitor is running.

When no environment is specified, the monitor is run in the :default environment.

require 'resque'
Listerine::Monitor.new do
  name "Resque workers"
  description "This monitor makes sure that there is at least 1 Resque worker."
  environments :staging, :production
  assert do
    if current_environment == :staging
      url = "redis://staging:[email protected]:6379"
    else
      url = "redis://production:[email protected]:6379"
    end
    redis_info = URI.parse(url)
    Resque.redis = Redis.new(:host => redis_info.host, :port => redis_info.port, :password => redis_info.password)
    Resque.workers.length > 0
  end
end

# Run all the monitors
Listerine::Runner.instance.run

When using multiple environments, you can also define different recipients for the failure notification. Set a criticality level on the monitor using is, and when defining recipients in the configure block, indicate which recipients are for which criticality level. Criticality levels are arbitrary symbols that you define. In this example we'll use :critical but it could be whatever you want.

Listerine::Monitor.configure do
  from "[email protected]"
  # This is the default recipient
  notify "[email protected]"

  # When an alert fails that is of criticality level critical, notify [email protected]
  notify "[email protected]", :when => :critical
end

Listerine::Monitor.new do
  name "Resque workers"
  description "This monitor makes sure that there is at least 1 Resque worker."
  environments :staging, :production
  # This monitor is critical when running in the production environment
  is :critical, :in => :production
  # The criticality levels can be anything you want. It defaults to :default to notify the default recipient.
  is :foobar, :in => :staging
  assert do
    # Setup a different connection based on the current environment
    if current_environment == :staging
      url = "redis://staging:[email protected]:6379"
    else
      url = "redis://production:[email protected]:6379"
    end
    redis_info = URI.parse(url)
    Resque.redis = Redis.new(:host => redis_info.host, :port => redis_info.port, :password => redis_info.password)
    Resque.workers.length > 0
  end
end

# Run all the monitors
Listerine::Runner.instance.run

If a recipient is not declared for a criticality level, Listerine will use the default recipient.

Criticality levels can be set globally in the configure block so you can make all production monitors :critical, etc.

Note: You don't need to use multiple environments to set criticality levels. These are perfectly valid monitors:

Listerine::Monitor.configure do
 notify "[email protected]"
 notify "[email protected]", :when => :critical
end

# Emails [email protected] when it fails
Listerine::Monitor.new do
 name "Site online"
 is :critical
 assert do
   # Some code to check that the site is online
 end
end

# Emails [email protected] when it fails
Listerine::Monitor.new do
 name "Internal wiki online"
 assert do
   ...
 end
end

Customizing notification thresholds

You might not want to get notified the first time a monitor fails. When defining a monitor, you can define variables to notify_after some number of consecutive failures, and after you've received a notification, to then_notify_every x failures after that.

These options can be defined locally on a monitor, or globally set in the configure block. By default, both values are set to 1.

Listerine::Monitor.new do
  name "Cache online"
  description "This monitor connects to the cache and tries to set and then get a key."
  # Don't notify until there have been 2 consecutive failures
  notify_after 2
  # After 2 failures, only send a new notification every 3 failures
  then_notify_every 3
  assert do
    # Connect to cache
    cache = ...
    # Set key to value, then ensure that you can pull it from the cache
    cache.set("key", "value")
    cache.get("key") == "value"
  end
end

Adding custom actions for failures

When a monitor fails, you might want to take custom action. For example, you might want to reboot a machine after 5 consecutive failures. To do that, you can pass a block to if_failing, which will be yielded to with the current consecutive failure count.

Listerine also provides a wrapper around sending mail, Listerine::Mailer.mail(to, subject, body) if you want to add custom notifications.

Listerine::Monitor.new do
  name "Cache online"
  description "This monitor connects to the cache and tries to set and then get a key."
  assert do
    ...
  end

  # Reboot the cache instance if there are 5 consecutive failures
  if_failing do |failure_count|
    if failure_count == 5
      Listerine::Mailer.mail("[email protected]", "Rebooting the cache", "Cache failed 5 times, rebooting")
      system("ec2-reboot-instances i-1234567")
    end
  end
end

Configure Options

The following are the global options you can set in the configure block.

  • from - The email address from which your notifications are sent
  • notify with an optional :in => :environment - The recipients of notifications

You can set the follow options globally in the configure block to avoid having to redefine on each monitor:

  • is - defaults to :default
  • notify_after - defaults to 1
  • then_notify_every - defaults to 1

Helper functions

Because a common use case we have is to check if a website is online, there is a simple helper assert_online which creates an assert block to ensure that a website is returning a 200 status code.

Listerine::Monitor.new do
  name "Site online"
  is :critical
  description "Makes sure that the site is online."
  assert_online "http://www.example.com"
end

We use CloudFlare, and CloudFlare is somewhat flaky, so you can pass in :ignore_502 => true to ignore 502 errors.

Notes

The name field must be unique across all defined monitors.

Persistence

Right now the persistence for these monitors is stored in a Sqlite3 database which is stored by default at ENV["HOME"]/listerine-default.db. You can customize the path to this database in the configure block:

Listerine::Monitor.configure do
  persistence :sqlite, :path => "/data/monitors/functional.db"
end

Other persistence backends can be created by adhering to the interface Listerine::Persistence::PersistenceLayer. Then pass your new persistence layer in via the :persistence option. Listerine will automatically find a class that is the #capitalized version of the option, so :persistence => :mysql will new up Listerine::Persistence::Mysql.

Listerine Server

Listerine comes with a Sinatra-based front end, Listerine::Server, for checking the latest status of your monitors and enabling/disabling them on a per-environment basis.

Check examples/config.ru for a functional example (including HTTP basic auth).

Listerine Server

Installation

Passenger

See Phusion's guide:

Apache: http://www.modrails.com/documentation/Users%20guide%20Apache.html#_deploying_a_rack_based_ruby_application

Nginx: http://www.modrails.com/documentation/Users%20guide%20Nginx.html#deploying_a_rack_app

Rack::URLMap

If you want to load Listerine on a subpath, possibly alongside other apps, it's easy to do with Rack's URLMap.

I don't really recommend this, since using a Sqlite backend, it means that your application server must also run the monitors, but you could conceptually do this if you want.

require 'listerine/server'

run Rack::URLMap.new \
  "/"       => Your::App.new,
  "/monitors" => Listerine::Server.new

Check examples/config.ru for a functional example (including HTTP basic auth).

Rails 3

You can also mount Listerine on a subpath in your existing Rails 3 app by adding require listerine/server to the top of your routes file or in an initializer then adding this to routes.rb.

mount Listerine::Server.new, :at => "/monitors"

Notes

The Listerine server will pick up any monitors that have any run history, meaning if you delete a monitor it will still show up. For now, simply delete your Sqlite database and it will be recreated the next time the monitors run.

Contributing

  1. Fork listerine
  2. Create a topic branch - git checkout -b my_branch
  3. Push to your branch - git push origin my_branch
  4. Create a Pull Request from your branch
  5. That's it!