get metrics from AWS ELBs

bin/AWScloudwatchELB.rb

@@ -0,0 +1,65 @@
+#!/usr/bin/env ruby
+## grab metrics from AWS cloudwatch 
+### David Lutz
+### 2012-07-15
+### gem install fog  --no-ri --no-rdoc
+
+require 'rubygems'
+require 'fog'
+require 'json'
+require '/opt/vacuumetrix/conf/config.rb'
+require '/opt/vacuumetrix/lib/Sendit.rb'
+
+if ARGV.length != 1
+        puts "I need one argument. The ELB name"
+        exit 1
+end
+
+dimensionId = ARGV[0]
+
+#AWS cloudwatch stats seem to be a minute or so behind
+
+startTime = Time.now.utc-180
+endTime  = Time.now.utc-120
+
+
+metricNames = {	"RequestCount"		=> "Sum",
+		"HealthyHostCount" 	=> "Minimum", 
+		"UnHealthyHostCount"	=> "Maximum",
+		"HTTPCode_ELB_5XX"	=> "Sum",
+		"HTTPCode_ELB_4XX"	=> "Sum",
+		"HTTPCode_Backend_2XX"	=> "Sum",
+		"HTTPCode_Backend_3XX"	=> "Sum",
+		"HTTPCode_Backend_4XX"	=> "Sum",
+		"HTTPCode_Backend_5XX"	=> "Sum"
+	}
+
+unit = 'Count'
+
+cloudwatch = Fog::AWS::CloudWatch.new(:aws_secret_access_key => $awssecretkey, :aws_access_key_id => $awsaccesskey)
+
+metricNames.each do |metricName, statistic|
+  response = cloudwatch.get_metric_statistics({
+           'Statistics' => statistic, 
+           'StartTime' =>  startTime.iso8601,
+           'EndTime'    => endTime.iso8601, 
+	   'Period'     => 60, 
+           'Unit'       => unit,
+	   'MetricName' => metricName, 
+	   'Namespace'  => 'AWS/ELB',
+	   'Dimensions' => [{
+	                'Name'  => 'LoadBalancerName', 
+	                'Value' => dimensionId 
+			}]
+           }).body['GetMetricStatisticsResult']['Datapoints']
+
+  metricpath = "AWScloudwatch.ELB." + dimensionId + "." + metricName 
+  begin
+  	metricvalue = response.first[statistic]
+  rescue
+	metricvalue = 0
+  end
+  metrictimestamp = endTime.to_i.to_s
+
+  Sendit metricpath, metricvalue, metrictimestamp
+end