Runtime

Node.js Development

We build services, workers and integration adapters on Node.js, where the workload is mostly waiting on other systems and concurrency is the point.

Usage

What we build with Node.js

Rationale

The workload fits the model
Integration work is mostly waiting on databases and third party APIs. An event loop handles thousands of concurrent waits without a thread each.
One language, one set of types
The worker, the API and the front end share validation schemas and domain types, which removes a whole category of mismatch.
Deployment is flexible
The same code runs as a serverless function, a container or a long lived process, so the hosting choice follows the workload.

Practice

Queue backed workers
Slow work leaves the request path. The handler acknowledges, enqueues, and a worker processes with retries and a dead letter queue.
Graceful shutdown
Processes stop accepting work, finish what is in flight and close connections, so a deployment does not drop a job midway.
Streaming for large payloads
Large files and exports stream rather than buffer, which keeps memory flat regardless of size.
Structured logging with correlation
Every log line carries a request or job identifier, so a failure can be traced across services rather than guessed at.

Related work

These projects share relevant engineering concerns. Each study identifies its documented technology choices.

Solutions

Tell us what you are building and what it has to connect to. We will come back with an approach, a scope and what it would take.