Adding a new column should be simple, but in production it carries weight. Schema changes affect query performance, application logic, and deployment safety. Done wrong, they cause downtime, block writes, or corrupt data. Done right, they improve capability without risk.
First, decide the column type and constraints. Nullability, default values, and indexing should be defined before the change hits production. Avoid adding non-null columns without defaults—they lock large tables during creation. If you need one, add it as nullable, backfill data, then enforce constraints.
Second, plan deployment steps. For large datasets, use online schema change tools or migration frameworks that avoid locking. Break the process into additive steps: create column, backfill, switch application code, remove old fields if needed.