-
Notifications
You must be signed in to change notification settings - Fork 3.8k
Description
Is there an existing issue for this?
- I have searched the existing issues
What happened?
Description
After it's version gate in #8175
it is no longer possible to use ec2SDConfigs with prometheus >= 3.8.0
But there seems to be nothing indicating that the old non-unified interface is removed.
The 3.8.0 release only references the adding of the new unified interface (link broken in release log)
I don't think we need to restrict users from using the ec2 interface, because prometheus cannot remove it in a minor version (unless they somehow considers it non-breaking). Particularly because we don't even seem to support the alternative aws sd config yet herein.
Steps to Reproduce
New prom, use the old interface. E.g. a ScrapeConfig with a spec like
spec:
ec2SDConfigs:
- filters:
- name: tag:Name
values:
- vault
port: 8200
refreshInterval: 1m
region: eu-west-1Expected Result
Prometheus operator should not reject it when it works upstream and remains documented.
Actual Result
Rejected resources
PrometheusOperatorRejectedResources
summary: Resources rejected by Prometheus operator
description: Prometheus operator in monitoring namespace rejected 2 prometheus/ScrapeConfig resources.
Prometheus Operator Version
relevant versions:
- image: quay.io/prometheus-operator/prometheus-operator:v0.88.0
- image: quay.io/prometheus/prometheus:v3.9.1Kubernetes Version
Server Version: v1.34.1-eks-3025e55Kubernetes Cluster Type
EKS
How did you deploy Prometheus-Operator?
helm chart:prometheus-community/kube-prometheus-stack
Manifests
prometheus-operator log output
ts=2026-01-21T13:10:52.144536719Z level=info caller=/go/pkg/mod/k8s.io/client-go@v0.34.3/tools/events/event_broadcaster.go:338 msg="Event occurred" object.name=vault-monitoring-scrapeconfig object.namespace=vault kind=ScrapeConfig apiVersion=monitoring.coreos.com/v1alpha1 type=Warning reason=InvalidConfiguration action=SelectingConfigurationResources note="\"vault/vault-monitoring-scrapeconfig\" was rejected due to invalid configuration: ec2SDConfigs: EC2 SD configuration is only supported for Prometheus version < 3.8.0. For Prometheus 3.8.0 onwards, please use AWS SD"Anything else?
No response