:plugin: ruby :type: filter /////////////////////////////////////////// START - GENERATED VARIABLES, DO NOT EDIT! /////////////////////////////////////////// :version: %VERSION% :release_date: %RELEASE_DATE% :changelog_url: %CHANGELOG_URL% :include_path: ../../../../logstash/docs/include /////////////////////////////////////////// END - GENERATED VARIABLES, DO NOT EDIT! /////////////////////////////////////////// [id="plugins-{type}s-{plugin}"] === Ruby filter plugin include::{include_path}/plugin_header.asciidoc[] ==== Description Execute ruby code. This filter accepts inline ruby code or a ruby file. The two options are mutually exclusive and have slightly different ways of working, which are described below. NOTE: This plugin's concurrency-safety depends on your code. Be sure to read up on <>. [id="plugins-{type}s-{plugin}-using-inline-script"] ===== Inline ruby code To add inline ruby in your filter, place all code in the `code` option. This code will be executed for every event the filter receives. You can also place ruby code in the `init` option. It will be executed only once during the plugin's register phase. For example, to cancel 90% of events, you can do this: [source,ruby] filter { ruby { # Cancel 90% of events code => "event.cancel if rand <= 0.90" } } If you need to create additional events, you must use a specific syntax `new_event_block.call(event)` like in this example duplicating the input event [source,ruby] filter { ruby { code => "new_event_block.call(event.clone)" } } NOTE: Defining methods in the <> can significantly reduce throughput. Use the <> instead. [id="plugins-{type}s-{plugin}-using-script-file"] ===== Using a Ruby script file As the inline code can become complex and hard to structure inside of a text string in `code`, it's then preferable to place the Ruby code in a .rb file, using the `path` option. [source,ruby] filter { ruby { # Cancel 90% of events path => "/etc/logstash/drop_percentage.rb" script_params => { "percentage" => 0.9 } } } The ruby script file should define the following methods: * `register(params)`: An optional register method that receives the key/value hash passed in the `script_params` configuration option * `filter(event)`: A mandatory Ruby method that accepts a Logstash event and must return an array of events Below is an example implementation of the `drop_percentage.rb` ruby script that drops a configurable percentage of events: [source,ruby] ---- # the value of `params` is the value of the hash passed to `script_params` # in the logstash configuration def register(params) @drop_percentage = params["percentage"] end # the filter method receives an event and must return a list of events. # Dropping an event means not including it in the return array, # while creating new ones only requires you to add a new instance of # LogStash::Event to the returned array def filter(event) if rand >= @drop_percentage return [event] else return [] # return empty array to cancel event end end ---- ===== Testing the ruby script To validate the behaviour of the `filter` method you implemented, the Ruby filter plugin provides an inline test framework where you can assert expectations. The tests you define will run when the pipeline is created and will prevent it from starting if a test fails. You can also verify if the tests pass using the logstash `-t` flag. For example above, you can write at the bottom of the `drop_percentage.rb` ruby script the following test: [source,ruby] ---- def register(params) # .. end def filter(event) # .. end test "drop percentage 100%" do parameters do { "percentage" => 1 } end in_event { { "message" => "hello" } } expect("drops the event") do |events| events.size == 0 end end ---- We can now test that the ruby script we're using is implemented correctly: [source,shell] ---- % bin/logstash -e "filter { ruby { path => '/etc/logstash/drop_percentage.rb' script_params => { 'drop_percentage' => 0.5 } } }" -t [2017-10-13T13:44:29,723][INFO ][logstash.filters.ruby.script] Test run complete {:script_path=>"/etc/logstash/drop_percentage.rb", :results=>{:passed=>1, :failed=>0, :errored=>0}} Configuration OK [2017-10-13T13:44:29,887][INFO ][logstash.runner ] Using config.test_and_exit mode. Config Validation Result: OK. Exiting Logstash ---- [id="plugins-{type}s-{plugin}-concurrency"] ==== Avoiding concurrency issues When events are flowing through a pipeline with multiple workers, a single shared instance of this filter may end up processing many events _simultaneously_. This means that your script needs to be written to avoid mutating shared state unless it is done in a thread-safe manner. In Ruby, the name of a variable determines its scope. The following guidance may help you avoid _accidentally_ mutating shared state: * Freely use Local Variables, whose name begins with a lower-case letter or an underscore (`_`). ** Local Variables are available only to the individual event being processed, and are automatically cleaned up. * Exercise caution when _modifying_ Instance Variables, whose names begin with `@` followed by a lower-case letter or an underscore (`_`). ** Instance Variables are shared between _all_ worker threads in this pipeline, which may be processing multiple events simultaneously. ** It is safe to _set_ Instance Variables in a <>-defined `register` function or with <>, but they should not be modified while processing events unless safe-guarded by mutual exclusion. ** Instance Variables are _not_ persisted across pipeline restarts or plugin crashes. * _Avoid_ using variables whose scope is not limited to the plugin instance, as they can cause hard-to-debug problems that span beyond the individual plugin or pipeline: ** Class Variables: begin with `@@`. ** Global Variables: begin with a `$`. ** Constants: begin with a capital letter. [id="plugins-{type}s-{plugin}-options"] ==== Ruby Filter Configuration Options This plugin supports the following configuration options plus the <> described later. [cols="<,<,<",options="header",] |======================================================================= |Setting |Input type|Required | <> |<>|No | <> |<>|No | <> |<>|No | <> |<>,{}|No | <> |<>,_rubyexception|No | <> |<>,_false|No |======================================================================= Also see <> for a list of options supported by all filter plugins.   [id="plugins-{type}s-{plugin}-code"] ===== `code` * Value type is <> * There is no default value for this setting. * This setting cannot be used together with `path`. The code to execute for every event. You will have an `event` variable available that is the event itself. See the <> for more information. [id="plugins-{type}s-{plugin}-init"] ===== `init` * Value type is <> * There is no default value for this setting. Any code to execute at logstash startup-time [id="plugins-{type}s-{plugin}-path"] ===== `path` * Value type is <> * There is no default value for this setting. * This setting cannot be used together with `code`. The path of the ruby script file that implements the `filter` method. [id="plugins-{type}s-{plugin}-script_params"] ===== `script_params` * Value type is <> * Default value is `{}` A key/value hash with parameters that are passed to the register method of your ruby script file defined in `path`. [id="plugins-{type}s-{plugin}-tag_on_exception"] ===== `tag_on_exception` * Value type is <> * Default value is `_rubyexception` Tag to add to events in case the ruby code (either inline or file based) causes an exception. [id="plugins-{type}s-{plugin}-tag_with_exception_message"] ===== `tag_with_exception_message` * Value type is <> * Default value is `false` If `true` adds a tag to the event that is the concatenation of `tag_with_exception_message` and the exception message. [id="plugins-{type}s-{plugin}-common-options"] include::{include_path}/{type}.asciidoc[]