Skip to main content

Module 1: Advanced Pod Scheduling

Module Information​

Difficulty: Intermediate Estimated Time: 90 minutes (15 min reading + 60 min lab + 15 min quiz)

What You Will Learn​

By completing this module, you will:

  • Master node affinity for placing workloads on specific nodes based on hardware or location requirements
  • Implement pod anti-affinity to spread replicas across nodes for high availability
  • Configure taints and tolerations to dedicate nodes for specific workloads
  • Combine scheduling strategies to achieve precise pod placement in production environments
  • Apply scheduling to the Voting App to optimize database performance and frontend availability

Prerequisites​

Before starting this module, you should have:

  • Completed Module 0: Introduction and Getting Started
  • A running KIND cluster with 3 worker nodes
  • The Example Voting App deployed and functional
  • Basic understanding of Kubernetes Deployments and Pods

Overview​

Your Voting App works, but pods land on random nodes. In production, you need control over where pods run. Databases should sit on fast storage, frontends need to spread across nodes for availability, and critical workloads may require dedicated hardware.

Right now, Kubernetes picks nodes randomly. A postgres pod might land on a node with slow disk I/O. All three vote replicas could end up on the same node, which becomes a single point of failure. If that node crashes, your entire voting frontend goes down.

This module teaches you to take control of the scheduler. You'll learn how to place postgres on SSD nodes, spread vote replicas across different machines, and use taints to keep general workloads off your database servers. These are the first production readiness improvements to your application.