How does the Node.js event loop work?

Node delegates I/O to the operating system and libuv, then processes completed work on a single JavaScript thread. Each loop iteration moves through phases. Timers run expired setTimeout and setInterval callbacks. The poll phase waits for I/O and runs ready callbacks. The check phase runs setImmediate callbacks. Close callbacks handle teardown events.

console.log("start");

setTimeout(() => console.log("timeout"), 0);
setImmediate(() => console.log("immediate"));
Promise.resolve().then(() => console.log("promise"));
process.nextTick(() => console.log("nextTick"));

console.log("end");

Microtasks versus phases

• The nextTick queue is drained before the promise microtask queue
• Microtasks run after the current operation and between event loop phases
• setTimeout with zero delay is still a timer callback and runs in a later iteration
• setImmediate runs in the check phase, which is why its order against timers is not guaranteed in the main module
• Long synchronous work blocks every phase and delays all callbacks

This design lets one thread keep many connections in flight, because the expensive waiting happens in the OS or the libuv thread pool. It does not make CPU work concurrent. A tight loop blocks timers, I/O callbacks and new requests until it finishes.