Hello,
I'm using filebeat (6.4.3) to ship Apache logs (and some system ones) directly to ES.
Currently Filebeat is configured by using apache2 module and wildcard path setting, pointing to a directory with near 16 000 log files (not all of them are constant updating).
The event intensity is about 1000/s at max.
Filebeat is constantly using more than 100% cpu (more than 1 core) on a "Intel(R) Xeon(R) CPU E5-2630 0 @ 2.30GHz" cpu.
My current registry file is of 1,5MB size (don't know if that matters).
This is a sample snippet from my filebeat log :
===================
2019-02-12T14:18:39.945+0200 INFO [monitoring] log/log.go:141 Non-zero metrics in the last 30s {"monitoring": {"metrics": {"beat":{"cpu":{"system":{"ticks":102890,"time":{"ms":6569}},"total":{"ticks":997290,"time":{"ms":48492},"value":997290},"user":{"ticks":894400,"time":{"ms":41923}}},"info":{"ephemeral_id":"707a2420-9c67-4a21-b1cc-605a5aa74d78","uptime":{"ms":1450532}},"memstats":{"gc_next":70045040,"memory_alloc":41175592,"memory_total":83141794256,"rss":-7925760}},"filebeat":{"events":{"active":7768,"added":16776,"done":9008},"harvester":{"open_files":617,"running":701,"started":114},"input":{"log":{"files":{"renamed":48}}}},"libbeat":{"config":{"module":{"running":0}},"output":{"events":{"acked":10158,"active":73,"batches":272,"failed":14,"total":10245},"read":{"bytes":216823,"errors":3},"write":{"bytes":8552477}},"pipeline":{"clients":20,"events":{"active":4122,"filtered":6364,"published":10174,"retry":64,"total":16544},"queue":{"acked":6112}}},"registrar":{"states":{"current":7239,"update":9008},"writes":{"success":1130,"total":1130}},"system":{"load":{"1":9.57,"15":11.48,"5":10.86,"norm":{"1":0.3988,"15":0.4783,"5":0.4525}}}}}}
===================
I've also tried to do httprof profiling and I'm attaching the result of 30s profile (as a png).