A new column sounds small. In a schema, it can be simple or dangerous. Done wrong, it triggers locks, slows writes, and burns CPU. Done right, it is invisible to the user and painless to the system. The difference is in the approach.
Start by knowing the database engine. In PostgreSQL, adding a nullable column without a default is instant. Adding a column with a default on a huge table is not. MySQL behaves differently depending on version and storage type. Many cloud-managed databases roll schema changes into online operations, but the details still matter.
For high-traffic systems, use safe deploy patterns. Add the column as nullable or with a lightweight default. Backfill data in small batches. Once populated, add constraints or indexes in a separate migration. Each step isolates changes, reduces lock times, and keeps latency steady.