Part 4: Run It on AWS

The rooms, the middleware and the invitations don’t know where they run. In this part, they move to AWS: Postgres on RDS, runners on ECS Fargate, and the API on Lambda.

Finish AWS setup and install @alchemy.run/aws.

Swap the database and the pool

src/Cluster.ts

import * as ClusterDocker from "@alchemy.run/effect-cluster/Docker";
import * as ClusterAWS from "@alchemy.run/effect-cluster/AWS";
import * as Layer from "effect/Layer";

export const Database = ClusterDocker.Database("Cluster");
export const Database = ClusterAWS.Database("Cluster");

export const Default = ClusterDocker.SwarmPool("Default", { replicas: 2 }).pipe(
export const Default = ClusterAWS.EcsPool("Default", { replicas: 2 }).pipe(
  Layer.provide(Database),
);

The database is now an RDS instance in a VPC of its own, and the pool an ECS service in that VPC, behind a load balancer that serves its gateway.

Serve the API from Lambda

src/Api.ts

import * as Cluster from "@alchemy.run/effect-cluster";
import * as Docker from "@alchemy.run/docker";
import * as Lambda from "@alchemy.run/aws/Lambda";
import * as Effect from "effect/Effect";
import { Database } from "./Cluster.ts";

export default class Api extends Docker.Service<Api>()(
export default class Api extends Lambda.Function<Api>()(
  "Api",
  Effect.gen(function* () {
    const database = yield* Cluster.Database.Database;
    return {
      main: import.meta.url,
      port: 3000,
      ports: [{ external: 8080, internal: 3000 }],
      networks: database.network?.kind === "docker" ? [{ name: database.network.name }] : [],
    };
  }).pipe(Effect.provide(Database)),
  { main: import.meta.url, functionUrl: true },

The Lambda function isn’t in the VPC: it calls the rooms through the gateway’s load balancer, with the token Alchemy binds into it.

Swap the providers

alchemy.run.ts

import * as Alchemy from "alchemy";
import * as AWS from "@alchemy.run/aws";
import * as Cluster from "@alchemy.run/effect-cluster";
import * as Docker from "@alchemy.run/docker";

  { providers: Layer.mergeAll(Docker.providers(), Cluster.providers()) },
  { providers: Layer.mergeAll(AWS.providers(), Cluster.providers()) },
  Effect.gen(function* () {
    const api = yield* Api;
    return { api: api.name };
    return { url: api.functionUrl };

alchemy deploy
curl -H "x-member: sam" https://<function-url>/lobby
# {"members":1}

The first deploy takes a while: a new RDS instance takes 10 to 15 minutes. After that, deploys only replace what changed.

That’s the whole app on AWS. To run on another cloud, swap ClusterAWS for ClusterGCP, ClusterFly, ClusterHetzner, ClusterKubernetes or ClusterCloudflare; see the platform pages.