Class: LogStash::Inputs::Metrics

Inherits:
Base show all
Defined in:
lib/logstash/inputs/metrics.rb

Overview

The Metrics inputs is responable of registring itself to the collector. The collector class will periodically emits new snapshot of the system, The metrics need to take that information and transform it into a ‘Logstash::Event`, which can be consumed by the shipper and send to Elasticsearch

Constant Summary

Constants included from Config::Mixin

Config::Mixin::ENV_PLACEHOLDER_REGEX, Config::Mixin::PLUGIN_VERSION_0_9_0, Config::Mixin::PLUGIN_VERSION_1_0_0

Constants inherited from Plugin

Plugin::NL

Instance Attribute Summary

Attributes inherited from Base

#params, #threadable

Attributes included from Config::Mixin

#config, #original_params

Attributes inherited from Plugin

#params

Instance Method Summary collapse

Methods inherited from Base

#do_stop, #initialize, plugin_type, #stop?, #tag

Methods included from Config::Mixin

#config_init, included, #replace_env_placeholders

Methods included from Util::Loggable

included, #logger

Methods inherited from Plugin

#close, #config_name, #debug_info, declare_plugin, #do_close, #eql?, #hash, #id, #initialize, #inspect, lookup, #metric, #metric=, #to_s

Constructor Details

This class inherits a constructor from LogStash::Inputs::Base

Instance Method Details

#register ⇒ Object 



16
17
 
# File 'lib/logstash/inputs/metrics.rb', line 16

def register
end

#run(queue) ⇒ Object 



19
20
21
22
23
24
25
26
27
28
29
 
# File 'lib/logstash/inputs/metrics.rb', line 19

def run(queue)
  @logger.debug("Metric: input started")
  @queue = queue

  # we register to the collector after receiving the pipeline queue
  metric.collector.add_observer(self)

  # Keep this plugin thread alive,
  # until we shutdown the metric pipeline
  sleep(1) while !stop?
end

#stop ⇒ Object 



31
32
33
34
 
# File 'lib/logstash/inputs/metrics.rb', line 31

def stop
  @logger.debug("Metrics input: stopped")
  metric.collector.delete_observer(self)
end

#update(snapshot) ⇒ Object 



36
37
38
39
40
41
42
43
44
45
 
# File 'lib/logstash/inputs/metrics.rb', line 36

def update(snapshot)
  @logger.debug("Metrics input: received a new snapshot", :created_at => snapshot.created_at, :snapshot => snapshot, :event => snapshot.metric_store.to_event) if @logger.debug?

  # The back pressure is handled in the collector's
  # scheduled task (running into his own thread) if something append to one of the listener it will
  # will timeout. In a sane pipeline, with a low traffic of events it shouldn't be a problems.
  snapshot.metric_store.each do |metric|
    @queue << LogStash::Event.new({ "@timestamp" => snapshot.created_at }.merge(metric.to_hash))
  end
end