Neon•2w ago

Recommended setup in long running node server

Hi all!
I have a node.js based backend running on railway serving a frontend,
and I'm confused regarding the driver choice and whether and where to use pooling.

My code thus far has been the following

import "dotenv/config";
import { drizzle } from "drizzle-orm/node-postgres";
import { Pool } from "pg";
import { env } from "./env";
import * as schema from "./schema";

const pool = new Pool({
    connectionString: env.DATABASE_URL,
});
export const db = drizzle(pool, { schema });

import "dotenv/config";
import { drizzle } from "drizzle-orm/node-postgres";
import { Pool } from "pg";
import { env } from "./env";
import * as schema from "./schema";

const pool = new Pool({
    connectionString: env.DATABASE_URL,
});
export const db = drizzle(pool, { schema });

However, the docs warn against double pooling, and I heard that server-side pooling might be better than client-side, especially when running replicas and the like.
I tried to use Client instead of Pool but it makes my backend crash on launch already.
Can you give me pointers towards a reliable production setup for a bog-standard fe/be webapp?

brilliant-limeOP•12/6/25, 8:33 PM

As might be obvious, I really dont know much about the intricacies of databases

instant-harlequin•12/6/25, 10:53 PM

Hey!
Definitely not obvious, the choice depends on a lot of things (especially in serverless applications)! Since you're running a persistent server, I'd recommend using and the direct connection string (no pgbouncer to avoid double pooling).

Iinstant-harlequin Hey! Definitely not obvious, the choice depends on a lot of things (especially i...

brilliant-limeOP•12/7/25, 12:52 AM

Will go with that then, seems like like simple and bulletproof solution. Thank you!