This is the Ruby adapter for the Raygun error reporter, https://raygun.com.
Add to your application's Gemfile:
$ bundle add raygun4ruby
And then execute:
$ bundle install
Run:
rails g raygun:install YOUR_API_KEY_HERE
You can find your API key in the Raygun app under "Application Settings".
You can then test your Raygun integration by running:
rails raygun:test
You should see an "ItWorksException" appear in your Raygun Crash Reporting dashboard. You're ready to zap those errors! ⚡
The generator will create a file in config/initializers called "raygun.rb". If you need to do any further configuration or customization of Raygun, that's the place to do it!
By default the Rails integration is set to only report Exceptions in the production environment. To change this behaviour, set config.enable_reporting to something else in config/initializers/raygun.rb.
Raygun4Ruby currently supports Rails 6 or later running on Ruby >= 3.0. If you're using an older version of Ruby or Rails please use an older version of Raygun4Ruby
To enable exception tracking in Sinatra, just add configure Raygun and use the Rack middleware in your app:
require 'raygun4ruby'
Raygun.setup do |config|
config.api_key = "YOUR_API_KEY_HERE"
end
use Raygun::Middleware::RackExceptionInterceptorrequire 'rubygems'
require 'raygun4ruby'
Raygun.setup do |config|
config.api_key = "YOUR_RAYGUN_API_KEY"
config.filter_parameters = [ :password, :card_number, :cvv ] # don't forget to filter out sensitive parameters
config.enable_reporting = Rails.env.production? # true to send errors, false to not log
end
begin
# your lovely code here
rescue => e
Raygun.track_exception(e)
end
# You may also pass a user object as the third argument to allow affected customers, like so
begin
# your lovely code here
rescue => e
# The second argument is the request environment variables
Raygun.track_exception(e, {}, user)
endYou can also pass a Hash as the second parameter to track_exception. It should look like a Rack Env Hash
If you'd like to customize how parameters are filtered, you can pass a Proc to filter_parameters. Raygun4Ruby will yield the params hash to the block, and the return value will be sent along with your error.
Raygun.setup do |config|
config.api_key = "YOUR_RAYGUN_API_KEY"
config.filter_parameters do |params|
params.slice("only", "a", "few", "keys") # note that Hash#slice is in ActiveSupport
end
endBreadcrumbs let you provide logging points in your code that will be collected and sent along with any exception sent to Raygun. This lets you have a better understanding of the events that happened in the system that lead up to the exception.
- Include it as a module in your class
class SomeClass
include Raygun::Breadcrumbs
def some_method
record_breadcrumb(
message: "<log message goes here>",
category: "some category to group them by, maybe authentication or external-apis for example",
level: :info, # or debug or warning etc, you can configure what level will get sent
metadata: {custom_data: 'can go here'},
)
end
endThis has the added benefit of recording the class the breadcrumb was recorded from automatically
- Call the
record_breadcrumbmethod manually
def some_method
Raygun.record_breadcrumb(
message: "<log message goes here>",
category: "some category to group them by, maybe authentication or external-apis for example",
level: :info, # or debug or warning etc, you can configure what level will get sent
metadata: {custom_data: 'can go here'},
# You can also set the class the breadcrumb was logged from
# It will only be set automatically using the included module approach
# Method and line number will get added automatically
class_name: self.class.name
)
endIf you are using Sinatra or another rack framework you will need to include the Breadcrumbs middleware, this is used for storing the breadcrumbs during a request
use Raygun::Middleware::BreadcrumbsStoreInitializer
(this must be before you use the Raygun::Middleware::RackExceptionInterceptor)
If you are using a non web based Ruby application you will have to call Raygun::Breadcrumbs::Store.initialize during your applications boot process. The store is per thread, but I have not tested it in a multi threaded application.
As an alternative to the above, you can also opt-in to the keys/values to be sent to Raygun by providing a specific whitelist of the keys you want to transmit.
This disables the blacklist filtering above (filter_parameters), and is applied to the entire payload (error, request, environment and custom data included), not just the request parameters.
In order to opt-in to this feature, set filter_payload_with_whitelist to true, and specify a shape of what keys you want (the default is below which is to allow everything through, this also means that the query parameters filtered out by default like password, creditcard etc will not be unless changed):
Raygun.setup do |config|
config.api_key = "YOUR_RAYGUN_API_KEY"
config.filter_payload_with_whitelist = true
config.whitelist_payload_shape = {
machineName: true,
version: true,
error: true,
userCustomData: true,
tags: true,
request: {
hostName: true,
url: true,
httpMethod: true,
iPAddress: true,
queryString: true,
headers: true,
form: {}, # Set to empty hash so that it doesn't just filter out the whole thing, but instead filters out each individual param
rawData: true
}
}
endAlternatively, provide a Proc to filter the payload using your own logic:
Raygun.setup do |config|
config.api_key = "YOUR_RAYGUN_API_KEY"
config.filter_payload_with_whitelist = true
config.whitelist_payload_shape do |payload|
# Return the payload mutated into your desired form
payload
end
endCustom data can be added to track_exception by passing a custom_data key in the second parameter hash.
begin
# more lovely code
rescue Exception => e
Raygun.track_exception(e, custom_data: {my: 'custom data', goes: 'here'})
endCustom data can also be specified globally either by setting config.custom_data to a hash
Raygun.setup do |config|
config.api_key = "YOUR_RAYGUN_API_KEY"
config.custom_data = {custom_data: 'goes here'}
endor to a proc, which gets passed the exception and environment hash
Raygun.setup do |config|
config.api_key = "YOUR_RAYGUN_API_KEY"
config.custom_data do |e, env|
{message: e.message, server: env["SERVER_NAME"]}
end
endYou can ignore certain types of Exception using the ignore option in the setup block, like so:
Raygun.setup do |config|
config.api_key = "MY_SWEET_API_KEY"
config.ignore << ['MyApp::AnExceptionIDontCareAbout']
endThe following exceptions are ignored by default:
ActiveRecord::RecordNotFound
ActionController::RoutingError
ActionController::InvalidAuthenticityToken
ActionDispatch::ParamsParser::ParseError
CGI::Session::CookieStore::TamperedWithCookie
ActionController::UnknownAction
AbstractController::ActionNotFound
Mongoid::Errors::DocumentNotFound
You can see this here and unignore them if needed by doing the following:
Raygun.setup do |config|
config.api_key = "MY_SWEET_API_KEY"
config.ignore.delete('ActionController::InvalidAuthenticityToken')
endYou can pass proxy settings using the proxy_settings config option.
Raygun.setup do |config|
config.api_key = "MY_SWEET_API_KEY"
config.proxy_settings = { address: "localhost", port: 8888 }
endRaygun can track how many customers have been affected by an error.
By default, Raygun looks for a method called current_user on your controller, and it will populate the customer's information based on a default method name mapping.
(e.g Raygun will call email to populate the customer's email, and first_name for the customer's first name)
You can inspect and customize this mapping using config.affected_user_mapping, like so:
Raygun.setup do |config|
config.api_key = "MY_SWEET_API_KEY"
config.affected_user_method = :my_current_user # `current_user` by default
# To augment the defaults with your unique methods you can do the following
config.affected_user_mapping = Raygun::AffectedUser::DEFAULT_MAPPING.merge({
identifier: :some_custom_unique_identifier,
# If you set the key to a proc it will be passed the user object and you can construct the value your self
full_name: ->(user) { "#{user.first_name} #{user.last_name}" }
})
endTo see the defaults check out affected_user.rb
If you're using Rails, most authentication systems will have this method set and you should be good to go.
The count of unique affected customers will appear on the error group in Raygun Crash Reporting. If your customer has an email attribute, and that email has a Gravatar associated with that address, you will also see your customer's avatar.
If you wish to keep it anonymous, you could set this identifier to something like SecureRandom.uuid and store that in a cookie, like so:
class ApplicationController < ActionController::Base
def raygun_user
cookies.permanent[:raygun_user_identifier] ||= SecureRandom.uuid
end
end(Remember to set affected_user_method to :raygun_user in your config block...)
Raygun can attach the version of your application to its error reports. In your Raygun.setup block, set version to the current version of your app.
Raygun.setup do |config|
config.version = "1.0.0.4" # you could also pull this from ENV or however you want to set it.
endTags can be added to track_exception by passing a tags key in the second parameter hash.
begin
# more lovely code
rescue Exception => e
Raygun.track_exception(e, tags: ['my', 'tags', 'go here'])
endTags can also be specified globally either by setting config.tags to an array
Raygun.setup do |config|
config.tags = ['heroku']
endor to a Proc, which gets passed the exception and environment hash. This proc must return an array of strings
Raygun.setup do |config|
config.api_key = "YOUR_RAYGUN_API_KEY"
config.tags do |e, env|
[env["SERVER_NAME"]]
end
endRaygun4Ruby also includes a Resque failure backend. You should include it inside your Resque initializer (usually something like config/initializers/load_resque.rb)
require 'resque/failure/multiple'
require 'resque/failure/raygun'
require 'resque/failure/redis'
Resque::Failure::Multiple.classes = [Resque::Failure::Redis, Resque::Failure::Raygun]
Resque::Failure.backend = Resque::Failure::MultipleRaygun4Ruby can track errors from Sidekiq (we test on 6+ but there's no reason it shouldn't work on older versions). All you need to do is add the line:
require 'raygun/sidekiq'Either in your Raygun initializer or wherever else takes your fancy :simple_smile:
To track affected customers, define a class method on your worker class that returns a user object.
Make sure the name of this method is the same as whatever you have defined as the affected_user_method in your Raygun configuration and that it returns an object that fits the mappings defined in affected_user_mapping
If you have not changed these, refer to Affected customers for the defaults
class FailingWorker
include Sidekiq::Worker
def perform(arg1, arg2)
end
# Your method must accept an array of arguments
# These will be the same as those passed to `perform`
def self.current_user(args)
arg1 = args[0]
arg2 = args[1]
user = User.find_by(name: arg1)
# Your method must return a user object
user
endBy default, Raygun4Ruby will unwrap Sidekiq::JobRetry::Handled exceptions and report the original error via Exception#cause. If you would prefer not to hear about retries, you can set config.track_retried_sidekiq_jobs to false in your Raygun configuration.
For a complete list of configuration options see the configuration.rb file
Oops! Just let us know by opening an Issue on Github.
- Fork the repo, and install dependencies (
bundle install) - Create your feature branch and write your code (
git checkout -b my-new-feature) - Write some tests, and make them pass (
bundle exec rake) - Commit your changes (
git commit -am 'Add some feature') - Push to the branch (
git push origin my-new-feature) - Create a new Pull Request. Thank you! ✨
- Build the gem (
gem build raygun4ruby.gemspec) - don't bother trying to build it on Windows, the resulting Gem won't work. - Install the gem (
gem install raygun4ruby-VERSION.gem)