Adding a new column sounds simple. In practice, it can break production if done wrong. Schema changes are dangerous because they touch the foundation of your data. Whether you work with PostgreSQL, MySQL, or a distributed SQL engine, the process matters.
First, understand your data size. A ALTER TABLE ADD COLUMN on a small dataset takes seconds. On tables with hundreds of millions of rows, it can lock writes for hours. This is why online schema change tools exist—pg_online_schema_change, gh-ost, or native ALTER TABLE features in newer PostgreSQL versions. They add the new column in a way that minimizes downtime.
Second, choose defaults carefully. Setting a non-null column with a default forces a rewrite of all rows. That rewrite is physical and expensive. A nullable column with no default adds instantly. You can backfill data later in batches, using indexed updates to control impact.