When you add a new column to a table, you change the shape of your data. It affects queries, indexes, caching, and downstream systems. The database engine must alter internal structures, and the impact depends on engine type, table size, and load patterns. In PostgreSQL, ALTER TABLE ADD COLUMN is fast if the column has no default. Adding a default will rewrite every row, locking writes until complete. In MySQL, large table alters can block reads and writes if not handled with tools like pt-online-schema-change or native online DDL.
A new column must be designed with purpose. Choose the right data type. Avoid implicit type casts that slow queries. Consider column order only when working with storage engines or formats where it matters. For large datasets, measure the storage impact before deployment.
Plan the rollout. Use database migrations in version control. Deploy the schema first, then backfill in batches. Add indexes after the backfill to avoid locking the table for the entire operation. Monitor replication lag, since long-running alters or backfills can stall read replicas.