Some better elasticsearch reliability.

- Added a configurable option for timing out ES HTTP requests.

 - Stop sending reporter messages after one message for one failure.

scripts/base/frameworks/logging/writers/elasticsearch.bro

@@ -17,6 +17,9 @@ export {
 	## e.g. prefix = "bro_" would create types of bro_dns, bro_software, etc.
 	const type_prefix = "" &redef;
 
+	## The time before an ElasticSearch transfer will timeout.
+	const transfer_timeout = 2secs;
+
 	## The batch size is the number of messages that will be queued up before 
 	## they are sent to be bulk indexed.
 	## Note: this is mainly a memory usage parameter.