database connection pooling: overlooked or essential?

hey everyone, been thinking a lot about connection pooling lately. We've implemented it in a few places, and its definitely smoothed out some performance bumps, especially under heavy load. But Im curious, for those of you working with distributed systems and microservices, how critical is database connection pooling in your architecture? Are there scenarios where youve seen it become a bottleneck, or are there alternative strategies you prefer for managing database connections efficiently?