Autoscaling

Overview 2 AM, phone buzzing with alerts. I open it up — 40+ pods in the TKE cluster suddenly enter Pending state, business APIs timing out across the board. A quick kubectl describe pod reveals the scheduling failure reason, crystal clear: Insufficient tke.cloud.tencent.com/eni-ip. This wasn’t some exotic technical puzzle. The root cause was straightforward: under VPC-CNI mode, the auxiliary IPs bound to the elastic network interfaces (ENIs) on the nodes were exhausted....

Overview Autoscaling is one of Kubernetes’ most attractive capabilities—scale up when traffic arrives, scale down when it leaves, ensuring service quality while controlling costs. But “automatic” doesn’t mean “mindless.” Poorly configured autoscaling can cause: delayed scaling leading to service degradation, aggressive scale-down disrupting long connections, and flapping scaling wasting resources. K8s autoscaling consists of three layers plus an event-driven extension: Layer Component Scaling Dimension Trigger Pod horizontal HPA Pod replica count CPU/memory/custom metrics Pod vertical VPA Pod resource quotas CPU/memory historical usage Node Cluster Autoscaler Node count Pending Pods Event-driven KEDA Pod replica count Event sources (Kafka/Redis/…) Based on Kubernetes v1....