You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Describe the bug a clear and concise description of what the bug is.
I'd like to use the serviceDiscoveryRole: EndpointSlice configuration in the kube-prometheus-stack, and therefore manually created EndpointSlice objects, identical to the Endpoints objects that are created by the kube-prometheus-stack when filling out the endpoints: section for a component, e.g. kubeControllerManager.
The EndpointSlice objects have the correct endpoint IP addresses and ports in them. However, they don't seem to be recognized by Prometheus, as the corresponding target in the Prometheus UI is empty and no metrics for the target are displayed.
I wanted to ask whether the discovery of EndpointSlice objects using the kube-prometheus-stack is already supported, and would be very glad if someone could help me with configuring this the right way.
What's your helm version?
v3.10.1
What's your kubectl version?
v1.26.0
Which chart?
kube-prometheus-stack
What's the chart version?
v62.3.1
What happened?
Set serviceDiscoveryRole: EndpointSlice and created EndpointSlice objects for kubeControllerManager and kubeScheduler with the correct IP addresses and ports. However, the EndpointSlice objects are not recognized by the Prometheus ServiceMonitors created when installing the kube-prometheus-stack, and metrics for these targets are missing (targets section in the Prometheus UI is empty as well).
What you expected to happen?
EndpointSlice objects should show up as targets in the Prometheus UI and metrics should be successfully scraped from them.
Describe the bug a clear and concise description of what the bug is.
I'd like to use the
serviceDiscoveryRole: EndpointSlice
configuration in the kube-prometheus-stack, and therefore manually created EndpointSlice objects, identical to the Endpoints objects that are created by the kube-prometheus-stack when filling out theendpoints:
section for a component, e.g. kubeControllerManager.The EndpointSlice objects have the correct endpoint IP addresses and ports in them. However, they don't seem to be recognized by Prometheus, as the corresponding target in the Prometheus UI is empty and no metrics for the target are displayed.
Support for the EndpointSlice role was introduced to Prometheus Operator in v0.76.0: prometheus-operator/prometheus-operator#3862
I wanted to ask whether the discovery of EndpointSlice objects using the kube-prometheus-stack is already supported, and would be very glad if someone could help me with configuring this the right way.
What's your helm version?
v3.10.1
What's your kubectl version?
v1.26.0
Which chart?
kube-prometheus-stack
What's the chart version?
v62.3.1
What happened?
Set
serviceDiscoveryRole: EndpointSlice
and created EndpointSlice objects for kubeControllerManager and kubeScheduler with the correct IP addresses and ports. However, the EndpointSlice objects are not recognized by the Prometheus ServiceMonitors created when installing the kube-prometheus-stack, and metrics for these targets are missing (targets section in the Prometheus UI is empty as well).What you expected to happen?
EndpointSlice objects should show up as targets in the Prometheus UI and metrics should be successfully scraped from them.
How to reproduce it?
No response
Enter the changed values of values.yaml?
prometheus:
prometheusSpec:
serviceDiscoveryRole: EndpointSlice
Enter the command that you execute and failing/misfunctioning.
helm install kube-prom-stack -f values.yaml -n monitoring --version 62.3.1 prometheus-community/kube-prometheus-stack
Anything else we need to know?
No response
The text was updated successfully, but these errors were encountered: