Skip to content

06-NESTJS-INTEGRATION.md

Latest commit

 

History

History
127 lines (102 loc) · 3.38 KB

06-NESTJS-INTEGRATION.md

File metadata and controls

127 lines (102 loc) · 3.38 KB

6. NestJS Integration

Wiring this database into a NestJS backend running on the same host, next to Elasticsearch and Redis.

Scope: this is an integration reference only. This repository provides the database — it contains no NestJS application code. The snippets below belong in your separate NestJS project; copy what you need. Nothing here needs to be added to this repo.

6.1 Connection details

Host:     127.0.0.1
Port:     15409
User:     local_dev
Password: <your .env value>
Database: local_db
URL:      postgresql://local_dev:<password>@127.0.0.1:15409/local_db

Put it in your NestJS app's .env (separate from this repo's .env):

DATABASE_URL=postgresql://local_dev:<password>@127.0.0.1:15409/local_db

6.2 With TypeORM

// app.module.ts
TypeOrmModule.forRoot({
  type: 'postgres',
  url: process.env.DATABASE_URL,
  autoLoadEntities: true,
  synchronize: false,          // use migrations, even in dev
  // Keep the pool small — Postgres max_connections is 50 and shared.
  extra: { max: 10, idleTimeoutMillis: 30000 },
});

6.3 With Prisma

// schema.prisma
datasource db {
  provider = "postgresql"
  url      = env("DATABASE_URL")
}
# Cap the pool so NestJS doesn't exhaust connections.
DATABASE_URL="postgresql://local_dev:<password>@127.0.0.1:15409/local_db?connection_limit=10&pool_timeout=20"

Enable extensions you use (e.g. pgvector) in a migration:

// already created by this repo, but Prisma can manage them too:
generator client {
  provider        = "prisma-client-js"
  previewFeatures = ["postgresqlExtensions"]
}
datasource db {
  provider   = "postgresql"
  url        = env("DATABASE_URL")
  extensions = [vector, pg_trgm, uuid_ossp(map: "uuid-ossp")]
}

6.4 Connection-pool sizing (important on a shared laptop)

max_connections = 50 is shared by all clients. If you run NestJS in watch mode (which may spawn extra connections), plus a migration tool, plus a GUI, you can exhaust it. Guidance:

  • Keep each app pool ≤ 10.
  • For many short-lived connections, add PgBouncer in front (transaction pooling). A drop-in service can be added to docker-compose.yml later.
  • Symptoms of exhaustion: FATAL: sorry, too many clients already. Fix: lower app pool size, or raise max_connections (costs RAM — see doc 5).

6.5 Running NestJS itself in Docker?

If your NestJS service is also a Compose service and you want it on the same Docker network, attach it to this repo's network and use the service name instead of 127.0.0.1:

# in your app's compose file
services:
  api:
    networks: [local_net]
    environment:
      DATABASE_URL: postgresql://local_dev:<password>@local-postgres:5432/local_db
networks:
  local_net:
    external: true
    name: local_net

Inside the Docker network the port is 5432 (container port), not 15409.

6.6 Vector search from NestJS (pgvector quick example)

CREATE TABLE documents (
  id        bigserial PRIMARY KEY,
  content   text,
  embedding vector(1536)
);
CREATE INDEX ON documents USING hnsw (embedding vector_cosine_ops);
// nearest neighbours
await dataSource.query(
  `SELECT id, content FROM documents ORDER BY embedding <=> $1 LIMIT 5`,
  [`[${embedding.join(',')}]`],
);

✅ Next: never lose data → 7. Backup & Restore »