3 posts tagged with "nodejs"

Yes. I know. Forking a process in Node.js is slow. Instead of crying about it, let's see how we can handle it!

Let's assume that you have a service in which you:

1. Accept a request
2. Fork a process with child_process.fork
3. Execute some code within that process
4. Exit from the child process
5. Complete the request

Probably the first thing you tried was to receive the request, spin up a process, do whatever you need in the processor, and exit. You timed the whole thing and your jaw dropped that it took a million years for the request to complete, even if you are just doing a console.log('I love kittens') inside your processor.

Don't bother. I will tell you right now that the bottleneck is the forking.

This is a post about a specific Node.js library, named bull, which is used to execute background jobs.

The reason I am writing this is to address some cases which the library does not cover out of the box and share our experience on how we solved them in case others have the same needs.

In one of our backend services, we have a class, that is basically the business model of the service and is used in many locations within the code. At some point, we decided that we need to do some caching since a few methods were doing very expensive calls to other remote services and they didn’t need to occur that often.

How do you add caching to a class that is used very often within the rest of the codebase in the least intrusive way? You wrap the prototype of the class of course! And how do you wrap the prototype of the class? With another function that does the wrapping of course!