Merge pull request #2379 from satnam6502/fluentd-ek-doc

Update logging documentation to reflect on by default logging

contrib/logging/fluentd-ek/Makefile

@@ -39,6 +39,12 @@ logger-up:
 logger-down:
 	-${KUBECTL} delete pods synthetic-logger-0.25lps-pod
 
+logger10-up:
+	-${KUBECTL} create -f synthetic_10lps.yml
+
+logger10-down:
+	-${KUBECTL} delete pods synthetic-logger-10lps-pod
+
 get:	
 	${KUBECTL} get pods
 	${KUBECTL} get services

contrib/logging/fluentd-ek/README.md

@@ -1,6 +1,6 @@
 # Logging Pods in a Kubernetes Cluster using Fluentd, Elasticsearch and Kibana
 
-To enable logging of the activity inside a [pod](../../../docs/pods.md) using Elasticsearch set ``FLUENTD_ELASTICSEARCH=true`` in ``config-default.sh`` before launching the cluster. When a cluster is created a Docker container will be placed on each node which uses the [Fluentd](http://www.fluentd.org/) log collector to shovel all the container logs to an instance of [Elasticsearch](http://www.elasticsearch.org/) (or anything else that will listen to [Logstash](http://logstash.net/docs/1.4.2/tutorials/getting-started-with-logstash) format JSON on port 9200).
+When a GCE Kubernetes cluster is created a [pod](../../../docs/pods.md) will be placed on each node which uses the [Fluentd](http://www.fluentd.org/) log collector to collect all the Docker container logs and send them to an instance of [Elasticsearch](http://www.elasticsearch.org/) (or anything else that will listen to [Logstash](http://logstash.net/docs/1.4.2/tutorials/getting-started-with-logstash) format JSON on port 9200).
 
 We can verify that a Fluentd collector is running by ssh-ing into one of the nodes and looking at the running containers.
 

contrib/logging/fluentd-ek/synthetic_10lps.yml

@@ -0,0 +1,29 @@
+# This pod specification creates an instance of a synthetic logger. The logger
+# is simply a program that writes out the hostname of the pod, a count which increments
+# by one on each iteration (to help notice missing log enteries) and the date using
+# a long format (RFC-3339) to nano-second precision. This program logs at a frequency
+# of 0.25 lines per second. The shellscript program is given directly to bash as -c argument
+# and could have been written out as:
+#   i="0"
+#   while true
+#   do
+#     echo -n "`hostname`: $i: "
+#     date --rfc-3339 ns
+#     sleep 4
+#     i=$[$i+1]
+#   done
+
+apiVersion: v1beta1
+kind: Pod
+id: synthetic-logger-10lps-pod
+desiredState:
+  manifest:
+    version: v1beta1
+    id: synth-logger-10lps
+    containers:
+      - name: synth-lgr
+        image: ubuntu:14.04
+        command: ["bash", "-c", "i=\"0\"; while true; do echo -n \"`hostname`: $i: \"; date --rfc-3339 ns; sleep 0.1; i=$[$i+1]; done"]
+labels:
+  name: synth-logging-source
+