Broken-pipe issue after running weblogic with monitoring-exporter on Openshift

[priwen]

WebLogic Monitoring Exporter Issue. We are having some sort of like broken-pipe issue after running weblogic with monitoring-exporter, after running the pods for 1-2 days it stops the metric feed and only by bouncing the pods only it resumes normal operations but only for a day or 2.

attached is the log:

csv-adminserver-monitoring-exporter (3).log

few things to consider:

1. Kubernetes platform is OpenShift
2. collectord is used to collect metric data that forward to splunk for scrapping (https://www.outcoldsolutions.com/docs/monitoring-openshift/)
3. application runs fine its just that metrics upload to splunk is unstable.
4. Pod qouta and limits resources is not being exhausted
5. Only experiencing when application is deployed if its an empty domain there's no broken-pipe issue error.

Any idea that can help identify where the RCA might be is welcome.

Thanks.