flapjack 0.6.32 → 0.6.33

data/doc/DEBUGGING.md

@@ -0,0 +1,28 @@
+Here's some useful ways of debugging various things in Flapjack.
+
+Redis Connection Leaks
+
+The following command reports every five seconds the number of connections to redis (as reported by lsof) and the totall number of EM external protocol connections currently in place:
+
+production paths:
+
+    while true ; do
+      echo -n "EM connection count: "
+      tail -50000 /var/log/flapjack/flapjack.log | grep -i "connection count" \
+      | tail -1 | awk '{ print $5 }'
+      echo -n "lsof redis: "
+      sudo lsof -p `cat /var/run/flapjack/flapjack.pid` | grep :6379 | wc -l
+      sleep 5
+    done
+
+development paths:
+
+    while true ; do
+      echo -n "EM connection count: "
+      tail -50000 log/flapjack.log | grep -i "connection count" \
+      | tail -1 | awk '{ print $5 }'
+      echo -n "lsof redis: "
+      sudo lsof -p `cat tmp/pids/flapjack.pid` | grep localhost:6379 | wc -l
+      sleep 5
+    done
+

data/lib/flapjack/api/entity_check_presenter.rb

@@ -27,23 +27,25 @@ module Flapjack
         num_states = hist_states.size
 
         hist_states.each_with_index do |obj, index|
-          if (index == 0)
+          ts = obj.delete(:timestamp)
+          if index == (num_states - 1)
+            # last (even if the only one)
+            obj[:start_time] = start_time ? [ts, start_time].max : ts
+            obj[:end_time]   = end_time
+          elsif (index == 0)
             # initial
-            ts = obj.delete(:timestamp)
             obj[:start_time] = start_time ? [ts, start_time].max : ts
             obj[:end_time]   = hist_states[index + 1][:timestamp]
-          elsif index == (num_states - 1)
-            # last
-            obj[:start_time] = obj.delete(:timestamp)
-            obj[:end_time]   = end_time
           else
             # except for first and last
-            obj[:start_time] = obj.delete(:timestamp)
+            obj[:start_time] = ts
             obj[:end_time]   = hist_states[index + 1][:timestamp]
           end
           obj[:duration] = obj[:end_time] ? (obj[:end_time] - obj[:start_time]) : nil
         end
 
+        # p hist_states
+
         hist_states.reject {|obj| obj[:state] == 'ok'}
       end
 
@@ -107,16 +109,18 @@ module Flapjack
             split_outs = []
 
             outs.each { |o|
-              next unless o[:start_time] < sm[:start_time] &&
-                o[:end_time] > sm[:end_time]
+              next unless o[:end_time] && (o[:start_time] < sm[:start_time]) &&
+                (o[:end_time] > sm[:end_time])
               o[:delete] = true
               split_outs += [{:state => o[:state],
                               :start_time => o[:start_time],
                               :end_time => sm[:start_time],
+                              :duration => sm[:start_time] - o[:start_time],
                               :summary => "#{o[:summary]} [split start]"},
                              {:state => o[:state],
                               :start_time => sm[:end_time],
                               :end_time => o[:end_time],
+                              :duration => o[:end_time] - sm[:end_time],
                               :summary => "#{o[:summary]} [split finish]"}]
             }
 
@@ -130,7 +134,7 @@ module Flapjack
           sched_maintenances.each do |sm|
 
             outs.each do |o|
-              next unless (sm[:start_time] < o[:end_time]) &&
+              next unless o[:end_time] && (sm[:start_time] < o[:end_time]) &&
                 (sm[:end_time] > o[:start_time])
 
               if sm[:start_time] <= o[:start_time] &&
@@ -142,9 +146,11 @@ module Flapjack
               elsif sm[:start_time] <= o[:start_time]
                 # partially overlapping on the earlier side
                 o[:start_time] = sm[:end_time]
+                o[:duration] = o[:end_time] - o[:start_time]
               elsif sm[:end_time] >= o[:end_time]
                 # partially overlapping on the later side
                 o[:end_time] = sm[:start_time]
+                o[:duration] = o[:end_time] - o[:start_time]
               end
             end
 
@@ -152,7 +158,7 @@ module Flapjack
           end
 
           total_secs = outs.inject(total_secs) {|ret, o|
-            ret[o[:state]] += (o[:end_time] - o[:start_time])
+            ret[o[:state]] += o[:duration] if o[:duration]
             ret
           }
 

data/lib/flapjack/notification/common.rb

@@ -14,7 +14,7 @@ module Flapjack
       def perform(notification)
         bootstrap
         @logger.debug "Woo, got a notification to send out: #{notification.inspect}"
-        dispatch(notification, :logger => @logger, :redis => ::Redis.new)
+        dispatch(notification, :logger => @logger, :redis => ::Resque.redis)
       end
 
     end

data/lib/flapjack/version.rb

@@ -1,5 +1,5 @@
 #!/usr/bin/env ruby
 
 module Flapjack
-  VERSION = "0.6.32"
+  VERSION = "0.6.33"
 end

data/spec/lib/flapjack/api/entity_check_presenter_spec.rb

@@ -72,6 +72,19 @@ describe 'Flapjack::API::EntityCheck::Presenter' do
     # TODO check the data in those hashes
   end
 
+  it "returns a (small) outage hash for a single state change" do
+    entity_check.should_receive(:historical_states).
+      with(nil, nil).and_return([{:state => 'critical', :timestamp => time - (4 * 60 * 60)}])
+    entity_check.should_receive(:historical_state_before).
+      with(time - (4 * 60 * 60)).and_return(nil)
+
+    ecp = Flapjack::API::EntityCheckPresenter.new(entity_check)
+    outages = ecp.outages(nil, nil)
+    outages.should_not be_nil
+    outages.should be_an(Array)
+    outages.should have(1).time_range
+  end
+
   it "a list of unscheduled maintenances for an entity check" do
     entity_check.should_receive(:maintenances).
       with(time - (12 * 60 * 60), time, :scheduled => false).and_return(maintenances)
@@ -120,7 +133,6 @@ describe 'Flapjack::API::EntityCheck::Presenter' do
     ecp = Flapjack::API::EntityCheckPresenter.new(entity_check)
     downtimes = ecp.downtime(time - (12 * 60 * 60), time)
 
-
     # 22 minutes, 3 + 8 + 11
     downtimes.should be_a(Hash)
     downtimes[:total_seconds].should == {'critical' => (22 * 60),
@@ -156,4 +168,30 @@ describe 'Flapjack::API::EntityCheck::Presenter' do
     downtimes[:downtime].should have(4).time_ranges
   end
 
+  it "returns downtime and handles an unfinished problem state" do
+    current = [{:state => 'critical', :timestamp => time - (4 * 60 * 60)},
+               {:state => 'ok',       :timestamp => time - (4 * 60 * 60) + (5 * 60)},
+               {:state => 'critical', :timestamp => time - (3 * 60 * 60)}]
+
+    entity_check.should_receive(:historical_states).
+      with(nil, nil).and_return(current)
+
+    entity_check.should_receive(:historical_state_before).
+      with(time - (4 * 60 * 60)).and_return(nil)
+
+    entity_check.should_receive(:maintenances).
+      with(nil, nil, :scheduled => true).and_return([])
+
+    ecp = Flapjack::API::EntityCheckPresenter.new(entity_check)
+    downtimes = ecp.downtime(nil, nil)
+
+    downtimes.should be_a(Hash)
+    downtimes[:total_seconds].should == {'critical' => (5 * 60)}
+    downtimes[:percentages].should == {'critical' => nil}
+    downtimes[:downtime].should be_an(Array)
+    # the last outage gets split by the intervening maintenance period,
+    # but the fully covered one gets removed.
+    downtimes[:downtime].should have(2).time_ranges
+  end
+
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: flapjack
 version: !ruby/object:Gem::Version
-  version: 0.6.32
+  version: 0.6.33
   prerelease: 
 platform: ruby
 authors:
@@ -11,7 +11,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2012-09-24 00:00:00.000000000 Z
+date: 2012-09-25 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: daemons
@@ -382,6 +382,7 @@ files:
 - dist/puppet/ruby/manifests/rubygems.pp
 - dist/puppet/sqlite3/manifests/dev.pp
 - doc/CONFIGURING.md
+- doc/DEBUGGING.md
 - doc/DEVELOPING.md
 - doc/GLOSSARY.md
 - doc/INSTALL.md