elasticsearch exporter used too much memory ,is it normal

[sunfloweryumi]

version：elasticsearch_exporter-1.5.0.linux
config ：elasticsearch_exporter-1.5.0.linux/elasticsearch_exporter --es.all --es.indices --es.cluster_settings --es.indices_settings --es.shards --es.snapshots --es.timeout=10s --es.uri xxx
my question: elasticsearch_exporter used 4.5G resident memory and 5G memory, is it normal?

[pierreblais]

Hi, I asking myself the same question. I'm currently running quay.io/prometheuscommunity/elasticsearch-exporter:v1.9.0 with:

resources:
  requests:
    cpu: 1
    memory: 3Gi
  limits:
    memory: 3Gi

And it often crashloop-ing. Which operation may consume as much resource in this metrics exporter ?

[sysadmind]

#1033 adds pprof profiling. This should allow pulling the memory breakdown from the running process. If you're running a build with that commit (currently only the master branch), you should be able to run the following while memory usage is high and post the PNG here (out.png) to help us track down the memory usage.

go tool pprof -png http://$elasticsearch_exporter_address:9114/debug/pprof/heap > out.png

[GhostLyrics]

#1033 adds pprof profiling. This should allow pulling the memory breakdown from the running process. If you're running a build with that commit (currently only the master branch), you should be able to run the following while memory usage is high […]

I'm also affected by the crash loops and was wondering when you were planning to ship a release of the exporter that includes
pprof profiling. Do you have any plans already?