I work a lot with Node.js backends and AWS Lambda, so the built-in connection pooling caught my attention. How does Prisma Postgres handle sudden spikes from serverless functions, and can we track connection or query usage per environment?
Every Prisma Postgres database has its own built-in PgBouncer pooler, so a burst of Lambdas shares a small set of real database connections instead of overwhelming it. For best results, create the client outside your handler and keep Lambda concurrency under your plan's limit (100 pooled connections on Starter, 500 on Pro, 1,000 on Business). You'll find the details in our connection pooling docs. Thanks for checking us out!