Hi Everyone, I am upgrading the Flink operator to ...
# troubleshooting
a
Hi Everyone, I am upgrading the Flink operator to 1.11, where in the CRD we can see a clear request and limit for resource allocation, which was missing in the older operator. Can someone please share the correct way to define the request and limit in the flink-resource.yml file with the operator 1.11? Currently, we are using the 1.4 operator and we configure the resource as resource_block under JM and TM like below.
Copy code
jobManager:
  replicas: {{ .Values.jobReplicas }}
  resource:
    {{- toYaml .Values.resources_blocks.jm | nindent 6 }}
taskManager:
  resource:
    {{- toYaml .Values.resources_blocks.tm | nindent 6 }}
Copy code
resources_blocks:
  jm:
    memory: 4G
    cpu: 2
  tm:
    memory: 8G
    cpu: 2
We observe that the CPU usage for the app, as indicated by the k8s pod metrics, is around 10%. Therefore, I am considering reducing the CPU for each JM and TM to a request of 0.5 and a limit of 1. Is there any reason not to change the resource configuration, particularly the CPU, in this manner? Or is there a minimum CPU value that each JM or TM must maintain? @Gyula F贸ra @Matyas Orhidi @Jeesmon Jacob can you please help and suggest Thanks in Advance
p
Hello there! I had pipelines running with
0.01
cpus sometimes. I'd say it's safe with you defining the growth factor
g
@Abhilash S.R please avoid tagging people directly in community help channels. This is completely voluntary from the contributors
馃憤 1