System Design
Message queues show up often in system design interviews—and in real systems—because they solve two big needs: decoupling components so they can scale independently, and doing asynchronous work without blocking the caller. When the interviewer asks how you’d handle “send an email after signup” or “process logs from millions of clients,” message queues are usually part of the answer. Here’s a clean way to talk about them.
A message queue sits between producers (who create work or events) and consumers (who process them). Producers don’t call consumers directly; they push into the queue. Consumers pull from the queue when they’re ready. So producers and consumers don’t depend on each other’s availability or speed. That decoupling means you can scale producers and consumers independently: add more consumers when the backlog grows, or add more producers when traffic spikes, without them being tightly wired together.