How to Implement Worker Threads in Node.js

Node.js runs JavaScript on a single thread. When you perform a CPU-intensive task like image resizing, cryptographic hashing, or processing a large dataset, the main thread is occupied for the duration. No other HTTP requests can be handled during this time, making your server unresponsive. Worker Threads solve this by running CPU-heavy code in separate threads with their own V8 instances.

Worker Threads run in the same process as the main thread and share the same memory space, which makes communication much faster. Child processes are fully separate operating system processes with their own memory, which provides stronger isolation but higher overhead. Use Worker Threads for CPU-intensive JavaScript work. Use Child Processes for running separate programs or scripts.

The worker_threads module is built into Node.js version 10.5 and above and does not require installation. Import Worker, isMainThread, parentPort, and workerData from the worker_threads module. These are the core primitives you need to create workers and communicate between them and the main thread.

The isMainThread boolean is true when the code runs in the main thread and false when it runs inside a Worker. Use this to write both the main thread logic and the worker logic in the same file. If isMainThread is true, create a new Worker pointing to the same file. If isMainThread is false, execute the CPU-intensive work and send the result back to the main thread.

When creating a new Worker, pass an options object with a workerData property containing the data the worker needs. This data is structured-cloned, meaning it is deeply copied into the worker's memory. The worker accesses this data through the workerData import from worker_threads. For large data sets, consider using SharedArrayBuffer to avoid the cost of copying.

Inside the worker thread, after completing the computation, call parentPort.postMessage with the result. The result is also structured-cloned back to the main thread. In the main thread, listen for the message event on the Worker instance to receive the result. The Worker acts like an EventEmitter with on, once, and off methods for event handling.

Listen for the error event on the Worker instance in the main thread to handle any uncaught exceptions thrown inside the worker. Listen for the exit event to detect when the worker finishes. The exit event receives an exit code where zero means normal completion and any non-zero code indicates an error. Always handle both events to prevent silent failures.

Creating a new Worker for every single task is expensive because each Worker has startup overhead. For applications that frequently need CPU work, create a pool of Workers that are reused across tasks. Maintain a fixed number of Workers and a queue of pending tasks. When a Worker finishes, assign it the next task from the queue immediately instead of shutting it down.