User performed Red Hat OpenShift Service on AWS (ROSA) version upgrade from v4.14.18 to v4.15.38 using OtelCollectorPod (splunk-otel-collector) version : v0.88.0. Splunk OTel Collector was installed using HELM and the install of the Collector is daemonset.
It was found that the receiver pod fails to start properly. Even after deleting and restarting the pod, it crashes with an "Error" → "Dump" state. On the Splunk APM side, active pods cannot be seen. However, data such as heap memory, which is collected by the agent, is visible.
1) Collect the Splunk Support Bundle as per the knowledge article Splunk OTel Collector and Kubernetes - Required Information to troubleshoot
2) Check the cluster receiver OTel Collector log file: pcn-stg-ns-splunktest/pods/splunk-otel-collector-1665567861-k8s-cluster-receiver-84749d8sj/otel-collector.log
3) Check the cluster receive log file shows the following error message
Error: failed to build pipelines: failed to create "kubeletstats" receiver for data type "metrics": cert path could not be read: open : no such file or directory
2024/12/06 02:54:21 main.go:94: application run finished with error: failed to build pipelines: failed to create "kubeletstats" receiver for data type "metrics": cert path could not be read: open : no such file or directoryThe kubeletstats receiver was incorrectly added to the cluster receiver configuration. The kubeletstats receiver should be configured within the OTEL agent daemonset only.
Kubeletstats is correctly configured in ./configmaps/splunk-otel-collector-1665567861-otel-agent/spec.yml and should not be configured in the cluster receiver ./configmaps/splunk-otel-collector-1665567861-otel-k8s-cluster-receiver/spec.yml.
Remove the kubeletstats configuration from teh cluster receiver configuration file.
| Revision | Publish Date | Comments |
|---|---|---|
1.0 |
29-Sep-2026
|
Initial Release |