Description
Component(s)
receiver/kubeletstats
What happened?
Description
We use Otel Collector version otel/opentelemetry-collector-contrib:0.120.0 to ingest metrics from Kubernetes via kubeletstatsreceiver into ClickHouse.
I've noticed that the results for the following metrics are identical:
k8s.node.cpu.usage - Total CPU usage (sum of all cores per second) averaged over the sample window
k8s.node.cpu.utilization - Node CPU utilization
select TimeUnix, MetricName, Value, ResourceAttributes, Attributes
from otel.otel_metrics_gauge
where
(MetricName = 'k8s.node.cpu.usage' OR MetricName = 'k8s.node.cpu.utilization')
AND ResourceAttributes['k8s.node.name'] = 'ip-10-31-49-125.eu-central-1.compute.internal'
AND TimeUnix >= NOW() - INTERVAL 120 SECOND
ORDER BY TimeUnix DESC;
another query with similar result
Steps to Reproduce
Install the Otel Collector using the official Helm chart (0.117.3).
Enable kubeletstatsreceiver for ingestion and activate the optional metric k8s.node.cpu.usage.
kubeletstats:
auth_type: serviceAccount
collection_interval: 10s
endpoint: ${env:K8S_NODE_IP}:10250
metric_groups:
- node
- pod
- container
- volume
node: '${env:K8S_NODE_NAME}'
k8s_api_config:
auth_type: serviceAccount
metrics:
k8s.node.cpu.usage:
enabled: true
Expected Result
k8s.node.cpu.utilization - provides CPU utilization in percentage.
Actual Result
k8s.node.cpu.utilization - provides the same result as k8s.node.cpu.usage.
Collector version
0.120.0
Environment information
Environment
EKS Cluster - Kubernetes 1.30