Kkhdevs·3d agoFour EKS versions and Karpenter 0.37If Karpenter on your EKS cluster logs "failed validating kubernetes version" with "error":"karpenter version is not compatible with K8s…
Kkhdevs·3d agoMinIO sat at 95% of its 512Mi limit for 30 days because it sizes itself from that limitIf your MinIO memory usage is high on Kubernetes ( kubectl top shows 495Mi against a 512Mi limit, an audit reports 7d memory working set…
Kkhdevs·3d agoOne failed IP assignment, 16 seconds after an EKS node was created, kept a counter at 1 for eight…If a pod on your EKS node got a FailedCreatePodSandBox event with add cmd: failed to assign an IP address to container, ipamd logged…
Kkhdevs·3d agoOne failed IP assignment, 16 seconds after an EKS node was created, kept a counter at 1 for eight…![The failed IP assignment on node-a at 12:32:40.926 UTC on a zoomed time axis: the line “DataStore has no available IP/Prefix addresses”…
Kkhdevs·4d ago30 days of Prometheus metrics sat on an emptyDir, and adding storageSpec would have deleted themIf Prometheus data was lost after a pod restart, or you are setting up kube-prometheus-stack persistent storage, the cause to check first…
Kkhdevs·4d agoEight namespaces and 23 CRDs that Argo CD could delete when a manifest left GitIf you searched “argocd deleted namespace”, “argocd prune deleted my namespace” or “argocd prevent namespace deletion”, run kubectl logs -n…
Kkhdevs·6d agoOur cluster had two nodes. Fifteen apps were allowed on only one of them.If your pods are not rescheduled when a node goes down and stay stuck Pending after node failure, run kubectl describe pod and read the…
Kkhdevs·6d agoSeven backup jobs succeeded. Five of them were on the database’s own disk.If your Kubernetes CronJob pg_dump backup writes to a PVC of the local-path StorageClass (the k3s default) and kubectl get jobs -A shows…A response icon1A response icon1
Kkhdevs·Nov 12Why did the kubelet evict pods for ephemeral-storage while 15 GiB were free?Because of evictionMinimumReclaim. Once free space crosses the kubelet's 5 % threshold, eviction continues until free space is above…A response icon1A response icon1
Kkhdevs·Nov 11Why did a pod that ran for months start crash-looping with “unrecognized option”?Because its first restart in six months pulled a different binary. The Deployment used coturn/coturn:latest with imagePullPolicy: Always…