while I was testing functionalists of Metricbeat following question came up on my mind. If the output is unreachable, can this cause partition overflow? Based on performed tests my conclusion was that data is stored locally during the unavailable output because as soon it became online all data was ingested to Elasticsaerch.
Metricbeat stores messages in its internal queue. By default it's an in-memory queue. It's configured to store 4096 messages. But you can change it to fit your environment.
Of course, this can cause overflows is you configure the internal queue. But if you use the defaults, I doubt that a regular server would fail.
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.