K8s Resource Planner

Understanding Kubernetes Resource Planning

Proper resource planning in Kubernetes is crucial for cluster stability, cost optimization, and application performance. By defining accurate requests and limits, you ensure the Kubernetes scheduler places your pods on appropriate nodes without wasting resources or causing Out-Of-Memory (OOM) kills.

📦 Requests vs Limits

• Requests: The minimum amount of CPU and memory guaranteed to a container. The scheduler uses this value to decide which node can host the pod.
• Limits: The maximum amount of CPU and memory a container is allowed to use. If a container exceeds its memory limit, it is terminated (OOMKilled). If it exceeds CPU, it is throttled.

🎯 QoS Classes

Kubernetes assigns Quality of Service (QoS) classes based on how you configure requests and limits:

• Guaranteed: Requests and limits are set and equal for all containers. Highest priority, last to be evicted.
• Burstable: At least one container has a request or limit set. Moderate priority.
• BestEffort: No requests or limits set. Lowest priority, first to be evicted under node pressure.

💡 Best Practices for Right-Sizing

1. Monitor Actual Usage: Use tools like Prometheus and Grafana to track actual CPU and memory consumption before setting limits.
2. Start with Requests: Always set requests first to ensure the scheduler can place your pods reliably.
3. Use Vertical Pod Autoscaler (VPA): Let VPA recommend optimal resource values based on historical usage.
4. Implement Horizontal Pod Autoscaler (HPA): Scale the number of replicas based on CPU/Memory utilization metrics.
5. Leave Headroom: Don't schedule pods to 100% node capacity. Leave 10-20% headroom for system daemons (kubelet, container runtime) and traffic spikes.

🔍 How to Use This Tool

1. Set your workload size: Adjust the number of pods/replicas your application requires.
2. Define per-pod resources: Input the CPU (in millicores) and Memory (in MiB) requested by a single pod.
3. Configure node specs: Enter the CPU cores and Memory (GiB) of your target worker nodes.
4. Review the plan: The tool calculates the total cluster resources needed, the recommended number of nodes, and the estimated utilization percentages.