i was building a starter with payloadcms , using supabase for postgres. there was a seed script running at build time, RSCs for every page that were using db queries under the hood.

after every build, the admin panel crashed with a connection pool error .

reason: vercel spins up multiple workers to build the app, and each worker acquires a connection from the pool . so all the workers were trying to get connections at the same time, and the pool got exhausted after the build.

so i used:

// payload.config.ts
 db: postgresAdapter({
    pool: {
      connectionString: process.env.DATABASE_URL || "",
      max: process.env.IS_PROD_BUILD === "true" ? 2 : 10,
      idleTimeoutMillis: 30_000,
      connectionTimeoutMillis: 10_000,
    },
    migrationDir: path.resolve(dirname, "migrations"),
    push: false,
  }),

// next.config.ts
export default (phase: string) => {
  const isBuildingForProduction = phase === PHASE_PRODUCTION_BUILD;
  const nextConfig: NextConfig = {
    env: {
      IS_PROD_BUILD: String(isBuildingForProduction),
    },
    images: {
      remotePatterns: [{ hostname: "images.unsplash.com" }],
    },
  };

  return withPayload(nextConfig);
};