Course contentsShow
← Cloud and Infrastructure
Lesson 298 of 2,9896. Auto ScalingFree lesson

Target Tracking Scaling Overview

Understanding target tracking policies that maintain a specific metric value by automatically adjusting capacity.

Target Tracking Scaling Overview

What you'll learn: How Auto Scaling can automatically adjust capacity to keep a specific metric at your desired value.


The Core Idea

Target tracking scaling is like cruise control for your cloud infrastructure. Just as cruise control maintains your car at a steady speed by automatically adjusting the throttle, target tracking maintains a specific metric (like CPU usage or request count) at a target value by automatically adding or removing instances.

Instead of writing complex rules like "if CPU > 70% add 2 instances," you simply tell Auto Scaling: "keep my average CPU at 50%." The system figures out the rest.

How It Works

You choose:

  1. A metric to track (e.g., average CPU utilization, ALB request count per target)
  2. A target value you want to maintain (e.g., 50% CPU)

Auto Scaling continuously monitors the metric. When it drifts above your target, Auto Scaling scales out (adds instances). When it falls below, Auto Scaling scales in (removes instances). The system calculates exactly how many instances to add or remove based on how far the metric deviates from your target.

Why It Matters

Target tracking is the simplest and most recommended scaling policy type. It handles both scaling out and scaling in automatically. You don't need to calculate thresholds or guess how many instances to add—the math happens behind the scenes.

This approach respects your minimum, maximum, and cooldown periods that you learned earlier, preventing wild oscillations while keeping your application performing smoothly.

Key Takeaway: Target tracking scaling automatically maintains a specific metric at your chosen target value by continuously adjusting capacity up or down, making it the easiest way to keep your application performing consistently.