05
Kubernetes Operations
Watch HPA autoscaling, rolling updates, and OOMKilled recovery in a 3-node cluster simulation.
IDLE
Scale100K
DAU
60K
Peak QPS
30K
WS Conns
10K
Data/Year
10 TB
Go Instances
6
Replicas
0 / 0
Version
v1
Updated
0
CPU
0%
QPS
0
v1 Running v2 Running Creating OOMKilled
K8s Operations
HPA
Horizontal Pod Autoscaler monitors CPU and custom metrics (QPS). When load exceeds target, it scales up pods. When load drops, it scales down after cooldown.
Rolling Update
Old pods are replaced one-by-one with new version. PodDisruptionBudget ensures minimum availability. Zero downtime if probes are configured correctly.
OOMKilled
When a container exceeds its memory limit, the kernel kills it. K8s restarts the pod automatically. If it keeps happening, you get CrashLoopBackOff.
Bin-packing
K8s scheduler places pods on nodes based on resource requests. Good bin-packing = higher utilization = lower cost. Three nodes show pod distribution in real time.