Horizontal vs Vertical Scaling
also called Scale Out vs Scale Up
Adding more machines versus making one machine bigger — and the fact that vertical is underrated for stateful tiers.
Vertical requires no application change, keeps everything in one process (no network, no distributed transactions, no partition key), and is bounded by the largest available instance — which is now very large indeed. Its downsides are a restart to resize, cost that rises faster than capacity at the top end, and no failure isolation.
Horizontal is effectively unbounded and gives redundancy, but requires the work to be partitionable and shifts complexity into the application: state must move out, load must be balanced, and coordination must be handled.
The practical guidance: scale stateless tiers horizontally, always — it is easy and it buys availability. Scale stateful tiers vertically first, because sharding a database is a one-way door and a bigger instance postpones it by years for a fraction of the engineering cost.