Implement ScoutApm::BackgroundJobIntegrations::GoodJob class

lib/scout_apm/background_job_integrations/good_job.rb

@@ -0,0 +1,50 @@
+module ScoutApm
+  module BackgroundJobIntegrations
+    class GoodJob
+      attr_reader :logger
+
+      def initialize(logger)
+        @logger = logger
+      end
+
+      def name
+        :good_job
+      end
+
+      def forking?
+        false
+      end
+
+      def install
+        require 'good_job'
+        GoodJob::ActiveRecordParentClass.class_eval do
+          include ScoutApm::Tracer
+
+          around_perform do |job, block|
+            # I have a sneaking suspicion there is a better way to handle Agent starting
+            # Maybe hook into GoodJob lifecycle events?
+            ScoutApm::Agent.instance.start_background_worker unless ScoutApm::Agent.instance.background_worker_running?
+            req = ScoutApm::RequestManager.lookup
+            latency = Time.now - job.scheduled_at rescue 0
+            req.annotate_request(queue_latency: latency)
+
+            begin
+              req.start_layer ScoutApm::Layer.new("Queue", job.queue_name)
+              started_queue = true # Following Convention
+              req.start_layer ScoutApm::Layer.new("Job", job.job_class)
+              started_job = true # Following Convention
+
+              block.call
+            rescue
+              req.error!
+              raise
+            ensure
+              req.stop_layer if started_job
+              req.stop_layer if started_queue
+            end
+          end
+        end
+      end
+    end
+  end
+end