Scalability lets an application absorb rising traffic, either by enlarging a single instance (vertical scalability) or by adding more machines (horizontal scalability). While vertical scaling suits monolithic databases, horizontal scaling powers modern web services. High availability pairs with horizontal scaling, deploying workloads across at least two Availability Zones to survive a data‑center outage.

AWS supplies four managed Elastic Load Balancers: the legacy Classic Load Balancer, the Application Load Balancer (layer 7), the Network Load Balancer (layer 4), and the newer Gateway Load Balancer (layer 3). They distribute traffic, perform health checks, terminate SSL/TLS, and support SNI for multiple certificates. Features like sticky sessions and cross‑zone balancing further improve resilience.

An Auto Scaling Group automates instance count based on CloudWatch alarms, ensuring a defined minimum and maximum. Policies range from simple target‑tracking of CPU to step‑scaling on custom metrics, and even predictive scaling that forecasts demand. Watching alarm thresholds and cooldown settings will reveal whether workloads stay responsive as traffic fluctuates.