How to modify the structure of a table?

0-2 years experience

1. 1How would you add a nullable `email` VARCHAR(255) column to an existing `users` table that already has 10k rows?
2. 2What happens if you try to change a `VARCHAR(50)` column to `VARCHAR(30)` when some existing rows have 40-character values?
3. 3You need to add an index on `created_at` for a reporting query. What's the exact command, and does it lock the table?

2-5 years experience

1. 1We're adding a `preferences` JSON column to a 50M-row `accounts` table for a new feature. Walk me through how you'd deploy this without downtime.
2. 2A migration to change a primary key from INT to BIGINT failed halfway through on a replica. How do you diagnose and recover?
3. 3You need to drop a `legacy_status` column that's referenced by a foreign key from an audit table. What steps do you take, and in what order?

5-8 years experience

1. 1Design a zero-downtime schema migration strategy for a 24/7 service with 100M+ row tables. Include tooling, validation, and rollback.
2. 2How do you coordinate schema changes across a sharded MySQL fleet where each shard has different data volumes?
3. 3Compare pt-online-schema-change, gh-ost, and native online DDL for a 500GB table. When would you pick each?

8+ years experience

1. 1We're splitting a monolith into microservices that share a MySQL cluster. How do you define schema ownership, migration approval, and breakage accountability across 10+ teams?
2. 2Design a schema versioning and rollback strategy for a multi-region active-active MySQL deployment with asynchronous replication.
3. 3How do you enforce schema change policies (code review, staging validation, canary deployment) across 50+ engineering teams without becoming a bottleneck?