Kubernetes Autoscaling Under Load

Simulate Pods, CPU spikes, HPA logic delays, cold starts, and node capacity constraint.

t = 0.0s

Speed Seed

Presets

Target CPU % 50%

Min Pods 2

Max Pods 16

HPA Sync Period 15s

Scale-Down Cooldown 120s

Scrape Interval 30s

Image Pull Time 8s

App Startup Delay 10s

Node Pod Capacity 3

Node Provision Delay 30s

Traffic Rate 150 rps

Ready Starting / Pulling Pending (No node) Terminating/Crashed Scraped CPU

Nodes

1 / 1

Pods

2 / 2

Avg CPU

75% (scr: 75%)

P95 Latency

15ms

Dropped RPS

0 rps

Stabilization

CPU Utilization (Real vs Scraped)

Replicas (Desired vs Ready vs Nodes)

Workload (RPS vs Dropped)

Node	Status	Usage / Cap

Pod	State	CPU / Mem

Interactive

Drag any entity to rearrange the topology canvas.

Scenarios

Trigger traffic spikes to observe the entire auto-scaling chain reaction!