Interactive VPA & Benchmarking

Experiment with the Vertical Pod Autoscaler and explore the hierarchy of resource controls to understand how Kubernetes manages workload benchmarking and resource allocation.

Interactive VPA Simulation

Play with the Vertical Pod Autoscaler to see how it works. Select a mode and watch how the VPA observes a running pod and makes recommendations. In "Auto" mode, you can trigger the update yourself!

VPA Control Panel

Pod Status

CPU Request: 250m

Memory Request: 256Mi

Live Workload Usage

Actual CPU: --

Actual Memory: --

VPA Recommendations

Rec. CPU: --

Rec. Memory: --

Select a mode to start the simulation.

VPA Benchmarking Patterns Explained

Explore the Resource Control Hierarchy

Cluster Level

ClusterResourceQuota

Node Level

Capacity / Taints

Namespace Level

ResourceQuota

Pod & Container Level

LimitRange

Container Spec

requests/limits

cpu, memory, pv

cpu, memory, pv

Interactive Pod Scheduling & VPA Flow

Hover over the steps to see how a pod moves from pending to running and how VPA provides recommendations based on the simulation above.

1. Pod Creation

A new pod is submitted with initial requests.

2. Scheduler Check

Checks if any node has enough resources to satisfy the pod's `requests`.

No → Pending

Pod remains `Pending` until resources are available.

Yes → Scheduled

Pod is assigned to a node and starts running.

3. VPA Monitors Usage

If VPA is active, it observes the pod's actual CPU and Memory usage over time.

4. VPA Provides Recommendations

Updates the VPA object with recommended `requests` values for optimal performance.