36 lines
1.2 KiB
YAML
36 lines
1.2 KiB
YAML
apiVersion: monitoring.coreos.com/v1
|
|
kind: PrometheusRule
|
|
metadata:
|
|
name: cilium-availability
|
|
namespace: prometheus
|
|
spec:
|
|
groups:
|
|
- name: cilium.availability
|
|
rules:
|
|
- alert: CiliumAgentDown
|
|
expr: |
|
|
kube_daemonset_status_desired_number_scheduled{namespace="kube-system", daemonset="cilium"}
|
|
-
|
|
kube_daemonset_status_number_ready{namespace="kube-system", daemonset="cilium"}
|
|
> 0
|
|
for: 5m
|
|
labels:
|
|
namespace: kube-system
|
|
severity: critical
|
|
annotations:
|
|
summary: One or more Cilium agents are unavailable
|
|
description: >-
|
|
Only {{ $value }} fewer Cilium agents are ready than desired.
|
|
Affected nodes may lose pod networking or service connectivity.
|
|
- alert: CiliumOperatorDown
|
|
expr: |
|
|
kube_deployment_status_replicas_available{namespace="kube-system", deployment="cilium-operator"} < 1
|
|
for: 5m
|
|
labels:
|
|
namespace: kube-system
|
|
severity: critical
|
|
annotations:
|
|
summary: The Cilium operator is unavailable
|
|
description: >-
|
|
No Cilium operator replica has been available for at least five minutes.
|