dont dive too deep into database indexes. when is enough, really enough?

hey everyone, i'm trying to get a better handle on database indexing strategies for system design interviews. we all know indexes are crucial for performance, but it feels like a slippery slope. too few, and queries crawl. too many, and writes become slow, and maintenance becomes a nightmare. what are your rules of thumb for deciding when you've indexed enough for a given table or service? are there specific query patterns or data volumes where you lean towards adding indexes versus optimizing queries or considering other solutions like denormalization or caching