Refactor logging to remove logger, use fluentd with kubernetes filter (#380)

Remove the `logger` container from the logging daemonset. 

Remove the outgoing call from the poolmgr to the logger.  Use Fluentd's Kubernetes filter to add function name and UID to influx metadata.  

This means fluentd now figures out when to start collecting function logs on its own, without being informed by poolmgr.  This is great for other execution strategies, and for autoscaling, where fission isn't in direct control of function pod creation.

Also adds an integration test to make sure logging keeps working.

logger/fluentd/fluent.conf

@@ -4,26 +4,32 @@
 
 <source>
   type tail
-  format json
+  format json 
   time_key time
-  path /var/log/fission/*.log
+  path "#{ENV['FLUENTD_PATH']}"
   time_format %Y-%m-%dT%H:%M:%S.%NZ
   tag fission.*
   read_from_head true
   refresh_interval 5
 </source>
 
+<filter fission.**>
+   type kubernetes_metadata
+</filter>
+
+
+<filter fission.**>
+   type flatten_hash
+   separator _
+</filter>
+
 <match fission.**>
-  type record_reformer
-  enable_ruby false
-  tag log
-  <record>
-    namespace ${tag_parts[4]}
-    pod       ${tag_parts[5]}
-    container ${tag_parts[6]}
-    funcname  ${tag_parts[7]}
-    funcuid   ${tag_parts[8]}
-  </record>
+   type record_reformer
+   enable_ruby false
+   tag log
+   <record>
+     funcuid ${kubernetes_labels_functionUid}
+   </record>
 </match>
 
 <match **>
@@ -35,7 +41,7 @@
     password  "#{ENV['INFLUXDB_PASSWD']}"
     use_ssl false
     time_precision s
-    tag_keys ["funcuid", "pod"]
+    tag_keys ["funcuid"]
     sequence_tag _seq
     buffer_type file
     buffer_path /var/log/fission/fluentd.buffer
@@ -45,4 +51,4 @@
     retry_limit 10
     retry_wait 1.0
     num_threads 2
-</match>
+</match>