Kibana version: 9.1.0, 9.1.7, 9.2.1

Elasticsearch version: dtto

Server OS version: wolfi image

Original install method (e.g. download page, yum, from source, etc.): ECK

Describe the bug:

Steps to reproduce:

1. install Kibana using ECK, declaring resource memory limit
2. observe memory growth untill the POD is OOMkilled.

Expected behavior:
There is active node.js garbage collector that keeps RAM usage reasonable

Screenshots (if relevant):

Any additional context:
I have checked node v8.getHeapStatistics() in fresh node process and the heap limit is 50% of memory limit.