Pattern 29 / SCALE

Auto-Scaling

Use this when matching capacity to variable traffic without manual intervention.

Pressure: Matching capacity to variable traffic without manual intervention
Mechanism: Observe metrics and change instance count according to scaling rules or forecasts
Toll: Scaling reacts with delay and can hide inefficient code or create cost surprises

Diagram for Auto-Scaling — Architecture plate29

Executive brief

Auto-Scaling fits when matching capacity to variable traffic without manual intervention. Mechanism: observe metrics and change instance count according to scaling rules or forecasts. Use it for flash sales, streaming events, cron bursts, and cloud-native workloads. The toll: scaling reacts with delay and can hide inefficient code or create cost surprises.

Use when

Flash sales, streaming events, cron bursts, and cloud-native workloads.

Example

Add api pods when p95 latency and cpu rise together.

Review framing

Describe the pressure first, then the mechanism, then the cost. That keeps the design grounded.

Same pressure family

Scaling Patterns

26Horizontal Scaling 27Vertical Scaling 28Load Balancing 30Database Connection Pooling