PVC 서비스 재시작이후 Pod 관련 문의입니다.

[#1]

PVC 서비스 재시작 후  'Pod Count' 서비스 재시작이전보다 감소하는 증상이 발생하여 정상유무를 확인하고자
문의를 드립니다.
+ 첨부 그림 파일처럼 서비스 재시작전에는 "Pod Cont" 갯수가 54개로 표시되지만, 서비스 재시작 후 18개로 줄어들어서 
  해당 증상이 비정상일 경우 확인해야 되는 항목과 참고할 수 있는 로그정보가 필요합니다.

'Pod Count'에는 어떤 정보가 포함되어 있나요?

[#2]

Here is the english translation: 

"I am writing to verify the status of the service after a PVC service restart, as the 'Pod Count' has decreased compared to the pre-restart level.
+ As shown in the attached image, the 'Pod Count' was 54 before the restart but dropped to 18 afterward. If this behavior is abnormal, please let me know what items I should check and what log information I should consult."

"What information is included in 'Pod Count'?"

[#3]

Hello @kiki_adm 

Thank you for reaching to Cloudera Community

Regarding the reported behavior where the Pod Count decreased from 54 to 18 after restarting the service, this does not necessarily indicate that pods were actually deleted or that there is an issue with the cluster.

The Pod Count displayed on the dashboard represents the number of pods being returned by the underlying monitoring/metrics query for the configured scope and filters. Therefore, the value can temporarily change after a service restart while the monitoring component is rebuilding or reloading its pod information.

To determine whether this is an actual pod-level issue or only a monitoring/dashboard discrepancy, we recommend first comparing the dashboard value with the actual Kubernetes pod count.

For example:

kubectl get pods -A --no-headers | wc -l

For a specific namespace:

kubectl get pods -n <namespace> --no-headers | wc -l
kubectl get pods -n <namespace>

It is also useful to check the pod status:

kubectl get pods -A

If Kubernetes shows that the expected number of pods are running while the dashboard shows only 18, this would indicate that the issue is likely related to metrics collection, monitoring, or the dashboard query, rather than an actual loss of pods.

Looking forward for your response

[#4]

Hello @kiki_adm 

To add to the answer already provided, I want to give some extra information. 

You asked this "What information is included in 'Pod Count'?"
This Grafana dashboard shows the pods that K8s has on the scope selected, either entire cluster or namespaces. 
In the screenshot we see some pods with the prefix impala, which means that those pods are for CDW (most likely). Those are for operations on CDW. 

During restart, those pods get stopped and not started until more tasks are executed on CDW. 

So the decrease on the number could be normal and expected. 

Hope this also helps to clarify your question.