As can be seen in the attached screenshot, at around 06:45:20 the heap scrapes the limit, when the GC count is 0 for the next minute.
This causes this node to be unresponsive for this time period.
It happens quite frequently from various nodes in the cluster (there are 9 data nodes)
I got the GC logs but can't quite understand what causes it, except for the long "Safepoint cleanup" starting at 03:45:29
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.