Adding a new column is one of the most common schema migrations. It sounds trivial. It can break production if done wrong. Columns store state. They change the contract between code and data. Every application reading from that table will see it. Every migration touches risk.
Before creating a new column, decide the data type. Choose NULL or NOT NULL with care. Understand default values and how they will be applied to existing rows. Avoid adding non-nullable columns to huge tables without defaults unless you want a long-lived lock. Know your database engine’s locking behavior. In MySQL, ALTER TABLE can block writes. In PostgreSQL, adding a nullable column is fast, but adding with DEFAULT rewrites the table.
Plan the rollout. Add the column first. Then backfill data in batches to avoid load spikes. Update application code after the column exists. Deploy in phases so consumers are ready before the data starts flowing. Watch for ORM-level caching or schema introspection issues.