Database connection pooling, is it essential or just overkill for small apps?

Hey everyone, I've been thinking a lot about database performance lately. For most of our microservices, connection pooling seems like a no-brainer. But for a really small, simple application that only gets a few requests a minute, is setting up and managing a connection pool still worth the overhead? Or are we better off just opening and closing connections per request in those edge cases? Curious to hear your experiences and recommendations.