fargate workload "Pod CPU Utilization" metric under prebuilt container Insight dashboard shows no data when the deployment pod does not have resource request/limit specified

[vasim38]

As per the implementation in [1] the "Pod CPU Utilization" metric for the fargate workload shows no data when the pod does not have resource request/limit specified.

However when the resource request/limit is specified, the "Pod CPU Utilization" uses the data received for the metric "pod_cpu_utilization_over_pod_limit" using ADOT collector.

---

apiVersion: v1
kind: Pod
metadata:
name: pod-with-resource-request-limit
namespace: default
labels:
app : pod-with-resource-request-limit
spec:
containers:
- name : test
image: nginx:1.14.2
resources:
requests:
cpu: "256m"
memory: "512Mi"
limits:
cpu: "512m"
memory: "1024Mi"

---

apiVersion: v1
kind: Pod
metadata:
name: pod-without-resource-request-limit
namespace: default
labels:
app : pod-without-resource-request-limit
spec:
containers:
- name : test
image: nginx:1.14.2

Validation :

kubectl exec -it pod-without-resource-request-limit sh
apt-get update -y
apt-get install sterss -y
stress --cpu 1 --timeout 300

kubectl exec -it pod-with-resource-request-limit sh
apt-get update -y
apt-get install sterss -y
stress --cpu 1 --timeout 300

The "Pod CPU Utilization" metric under prebuilt container Insight dashboard does not show data for the pod 'pod-without-resource-request-limit' but would show it for the pod 'pod-with-resource-request-limit'

If the above is expected, can you please confirm if there is a need to update the name for the metric "Pod CPU Utilization" when it uses the data from the "pod_cpu_utilization_over_pod_limit"

Appreciate the clarification here.

[1] https://aws-otel.github.io/docs/getting-started/container-insights/eks-fargate