registry-server creates high number of threads on Openshift 4

[arcmop]

/kind bug

We have noticed that the log server creates over than 900 threads inside the container on our nodes with (24 ~ 28) (44 ~ 48) cpus.
We also think it causes cpu saturation on the node.

Is it a normal condition?
Can it be improved with any parameter?

The content of database is:

The index.db is contained in "podman pull ibmcom/ibm-operator-catalog:latest" from https://hub.docker.com/r/ibmcom/ibm-operator-catalog/tags

I also attached de index.db on: https://github.com/arcmop/reporting-issues/raw/main/olm/index.7z

I would appreciate any support you could provide.