I'm trying to understand the following Filebeat INFO log:
INFO Non-zero metrics in the last 30s: filebeat.harvester.closed=2 filebeat.harvester.open_files=-1 filebeat.harvester.running=-2 libbeat.es.call_count.PublishEvents=138 libbeat.es.publish.read_bytes=2730237 libbeat.es.publish.write_bytes=555960458 libbeat.es.published_and_acked_events=635669 libbeat.publisher.published_events=545753 publish.events=917504 registrar.states.update=917504 registrar.writes=7
In particular, what do the following values mean? What are their differences? Which values are the important ones?
libbeat.es.published_and_acked_events: The events actually published, which is equal to the number of documents indexed successfully in Elasticsearch.
libbeat.publisher.published_events: If everything goes smooth, this should be equal to libbeat.es.published_and_acked_events. Otherwise, libbeat.publisher.published_events = libbeat.es.published_and_acked_events + #not_acked_events.
publish.events: The number of events in the original log files, including both the events being published, and the events excluded. Hence, if you drop/exclude some events/lines in Filebeat prospector, then publish.events = libbeat.publisher.published_events + #excluded_lines.
I am still trying to understand libbeat.es.publish.read_bytes and libbeat.es.publish.write_bytes. The size of my log files is 42GB. The size of the relevant lines in my log files is 37,508,554,927B (37.5GB). But after running expvar_rates.py, I get libbeat.es.publish.read_bytes: total 243121926 and libbeat.es.publish.write_bytes: total 56386818739, which do not match my data properties.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.