Adding a new column is trivial until it is not. Schema changes touch live systems. They break code paths you forgot existed. They slow queries you thought were fast enough. They cause deploys to stall while the database locks tables and holds back writes. In production, a single misstep can take the whole service down.
A new column should never be an afterthought. You need to plan. Decide if it allows nulls. Set defaults carefully. Know how it interacts with indexes. Think about rollbacks before you type ALTER TABLE. Large datasets might need an online schema change, run in batches to avoid blocking.
Coordinate with your application layer. Deploy code that can handle both the old and new schema. Use feature flags to control writes to the new column and to phase in reads. Test migrations against a copy of production data, not an empty dev database. Monitor latency, replication lag, and error rates during deployment.