Skip to main content
Sumo Logic

Kubernetes Alerts

To help determine if the Kubernetes cluster is available and performing well, the Sumo Logic monitors are provided with out of box alerts.

To help determine if the Kubernetes cluster is available and performing well, the Sumo Logic monitors are provided with out of box alerts.

Note: The alerts are built based on metrics datasets and have preset thresholds.

 

Name Description Trigger Type (Critical / Warning / MissingData) Alert Condition Recovery Condition
etcd Insufficient Members This alert is fired when we determine that etcd cluster has insufficient members. Critical >0 <=0
Kube API Down This alert is fired when KubeAPI disappears from Prometheus target discovery. Critical/MissingData <=0 >0
Kube Controller Manager Down This alert is fired when KubeControllerManager disappears from Prometheus target discovery. Critical <=0 >0
Kubelet Down This alert is fired when Kubelet disappears from Prometheus target discovery. Critical/MissingData <=0 >0
Kube Node Not Ready This alert is fired when a node is not ready. Critical/MissingData <=0 >0
Kube Scheduler Down This alert is fired when Kube Scheduler disappears from Prometheus target discovery. Critical/MissingData <=0 >0
Cluster CPU Utilization High This alert is fired when Cluster CPU utlization is high. Critical/Warning >0.90 <=0.90
Prometheus Remote Storage Failures This alert is fired when Prometheus fails to send samples to remote storage. Critical >1 <=1
Multiple Terminated Pods This alert is fired when we determine that there are pods that have been terminated. Critical >5 <=5
Pod Crash Looping This alert is fired when we determine that a pod is crash looping. Warning >0 <=0
Container Waiting This alert is fired when a pod container waiting longer than 1 hour. Warning >0 <=0
DaemonSet Not Scheduled This alert is fired when DaemonSet pods are not scheduled. Warning >0 <=0
DaemonSet Misscheduled This alert is fired when DaemonSet pods are miss-scheduled. Warning >0 <=0
StatefulSet Generation Mismatch This alert is fired when StatefulSet generation mismatch is determined due to possible roll-back. Warning >0 <=0
HPA Maxed Out This alert is fired when HPA is running at maximum replicas. Warning <=0 >0
Multiple Containers OOM Killed This alert is fired when multiple containers are OOM Killed. Warning >=5 <5