Call Stack vs Task Queue vs Microtask Queue in JS

Master the three core mechanisms of JavaScript's concurrency model. Covers the call stack execution model, macrotask queue scheduling, microtask queue priority, their interaction with the event loop, visualizing execution order, and debugging async timing issues.

JavaScriptadvanced
17 min read

JavaScript's single-threaded execution model relies on three interconnected mechanisms: the call stack processes synchronous code, the task queue holds deferred callbacks, and the microtask queue handles promise continuations with higher priority. Understanding how these interact is essential for predicting async execution order.

For the full event loop algorithm that coordinates these queues, see JavaScript Event Loop Internals Full Guide.

The Call Stack

javascriptjavascript
// The call stack is a LIFO (Last In, First Out) data structure
// that tracks function execution context
 
// CALL STACK MODEL
class CallStack {
  #frames = [];
  #maxDepth;
 
  constructor(maxDepth = 15000) {
    this.#maxDepth = maxDepth;
  }
 
  push(frame) {
    if (this.#frames.length >= this.#maxDepth) {
      throw new RangeError("Maximum call stack size exceeded");
    }
    this.#frames.push(frame);
    return this;
  }
 
  pop() {
    return this.#frames.pop();
  }
 
  peek() {
    return this.#frames[this.#frames.length - 1];
  }
 
  get depth() {
    return this.#frames.length;
  }
 
  get isEmpty() {
    return this.#frames.length === 0;
  }
 
  dump() {
    console.log("--- Call Stack ---");
    for (let i = this.#frames.length - 1; i >= 0; i--) {
      console.log(`  ${this.#frames[i].name} (${this.#frames[i].file}:${this.#frames[i].line})`);
    }
    console.log("------------------");
  }
}
 
// EXECUTION TRACE EXAMPLE
function main() {        // Stack: [main]
  const a = first();     // Stack: [main, first]
  const b = second();    // Stack: [main, second]
  return a + b;          // Stack: [main] -> returns, stack empty
}
 
function first() {       // Stack: [main, first]
  return helper(10);     // Stack: [main, first, helper]
}                        // helper returns -> [main, first] -> first returns -> [main]
 
function second() {      // Stack: [main, second]
  return 20;             // Stack: [main, second] -> second returns -> [main]
}
 
function helper(n) {     // Stack: [main, first, helper]
  return n * 2;          // Returns 20, pops helper
}
 
main(); // Returns 40
 
// STACK OVERFLOW
function infiniteRecursion() {
  return infiniteRecursion(); // RangeError: Maximum call stack size exceeded
}
 
// RUN-TO-COMPLETION
// Once a function starts, it runs until it returns (or throws)
// No other code can interrupt it on the same thread
// This is why long synchronous operations freeze the UI
 
function blockingOperation() {
  const start = Date.now();
  while (Date.now() - start < 5000) {
    // Blocks the call stack for 5 seconds
    // No callbacks, no rendering, no I/O during this time
  }
}

The Task Queue (Macrotask Queue)

javascriptjavascript
// The task queue holds callbacks from async APIs
// Only ONE task is dequeued per event loop iteration
 
// COMMON MACROTASK SOURCES:
// - setTimeout(callback, delay)
// - setInterval(callback, interval)
// - setImmediate(callback)          (Node.js)
// - I/O callbacks                   (fs.readFile, net.connect, etc.)
// - UI rendering events             (click, scroll, resize)
// - MessageChannel.port.onmessage
 
class TaskQueue {
  #queue = [];
 
  enqueue(task) {
    this.#queue.push({
      callback: task.callback,
      source: task.source,
      scheduledAt: Date.now()
    });
  }
 
  dequeueOne() {
    return this.#queue.shift() || null;
  }
 
  get length() {
    return this.#queue.length;
  }
 
  get isEmpty() {
    return this.#queue.length === 0;
  }
}
 
// TASK QUEUE BEHAVIOR
console.log("1: sync start");
 
setTimeout(() => console.log("2: timeout A"), 0);
setTimeout(() => console.log("3: timeout B"), 0);
setTimeout(() => console.log("4: timeout C"), 0);
 
console.log("5: sync end");
 
// OUTPUT:
// 1: sync start       (call stack - synchronous)
// 5: sync end         (call stack - synchronous)
// 2: timeout A        (task queue - first macrotask)
// [microtask checkpoint - empty]
// 3: timeout B        (task queue - second macrotask)
// [microtask checkpoint - empty]
// 4: timeout C        (task queue - third macrotask)
 
// EACH setTimeout callback is a SEPARATE macrotask
// Between each macrotask, ALL microtasks are drained
// This means a microtask scheduled inside timeout A
// runs BEFORE timeout B
 
setTimeout(() => {
  console.log("A: macrotask 1");
  Promise.resolve().then(() => console.log("B: microtask from A"));
}, 0);
 
setTimeout(() => {
  console.log("C: macrotask 2");
}, 0);
 
// OUTPUT:
// A: macrotask 1
// B: microtask from A    (drained before next macrotask)
// C: macrotask 2

The Microtask Queue

javascriptjavascript
// The microtask queue is drained COMPLETELY after each macrotask
// and after the initial script execution
 
// MICROTASK SOURCES:
// - Promise.prototype.then / catch / finally
// - queueMicrotask(callback)
// - MutationObserver (browser)
// - process.nextTick (Node.js, technically separate but similar priority)
 
class MicrotaskQueue {
  #queue = [];
 
  enqueue(microtask) {
    this.#queue.push({
      callback: microtask.callback,
      source: microtask.source
    });
  }
 
  // ALL microtasks are drained, including ones added during draining
  drainAll(callStack) {
    while (this.#queue.length > 0) {
      const task = this.#queue.shift();
      callStack.push({ name: task.source, file: "microtask", line: 0 });
      task.callback();
      callStack.pop();
    }
  }
 
  get length() {
    return this.#queue.length;
  }
}
 
// MICROTASK DRAINING DEMO
console.log("1: script start");
 
queueMicrotask(() => {
  console.log("2: microtask 1");
  queueMicrotask(() => {
    console.log("3: nested microtask");
    queueMicrotask(() => {
      console.log("4: deeply nested microtask");
    });
  });
});
 
queueMicrotask(() => console.log("5: microtask 2"));
 
console.log("6: script end");
 
// OUTPUT:
// 1: script start          (synchronous)
// 6: script end            (synchronous)
// 2: microtask 1           (first microtask)
// 5: microtask 2           (second microtask, queued before nested)
// 3: nested microtask      (queued during microtask 1)
// 4: deeply nested microtask (queued during nested microtask)
//
// ALL microtasks drain before ANY macrotask runs
// Nested microtasks are appended to the same queue and processed immediately
 
// STARVATION RISK
// Microtasks can prevent macrotasks from ever running
function microtaskStarvation() {
  let count = 0;
  function recurseMicrotask() {
    count++;
    if (count < 1_000_000) {
      queueMicrotask(recurseMicrotask);
    }
  }
  queueMicrotask(recurseMicrotask);
  setTimeout(() => console.log("This runs after 1M microtasks"), 0);
}

Complete Execution Model

javascriptjavascript
// FULL MODEL: Call Stack + Task Queue + Microtask Queue
 
class JSRuntime {
  #callStack = [];
  #taskQueue = [];
  #microtaskQueue = [];
 
  // Execute synchronous code
  execute(fn) {
    this.#callStack.push(fn.name || "anonymous");
    fn();
    this.#callStack.pop();
 
    // After each call stack empties, drain microtasks
    this.#drainMicrotasks();
  }
 
  // Schedule a macrotask
  scheduleTask(callback) {
    this.#taskQueue.push(callback);
  }
 
  // Schedule a microtask
  scheduleMicrotask(callback) {
    this.#microtaskQueue.push(callback);
  }
 
  // The event loop
  runEventLoop(maxIterations = 1000) {
    let iterations = 0;
 
    while (this.#taskQueue.length > 0 && iterations < maxIterations) {
      // Step 1: Pick ONE macrotask
      const task = this.#taskQueue.shift();
      this.#callStack.push("macrotask");
      task();
      this.#callStack.pop();
 
      // Step 2: Drain ALL microtasks
      this.#drainMicrotasks();
 
      iterations++;
    }
  }
 
  #drainMicrotasks() {
    while (this.#microtaskQueue.length > 0) {
      const micro = this.#microtaskQueue.shift();
      this.#callStack.push("microtask");
      micro();
      this.#callStack.pop();
    }
  }
}
 
// COMPREHENSIVE EXECUTION ORDER EXAMPLE
console.log("A: script start");                    // 1. sync
 
setTimeout(() => console.log("B: setTimeout 1"), 0);  // -> task queue
setTimeout(() => {                                    // -> task queue
  console.log("C: setTimeout 2");
  Promise.resolve().then(() => console.log("D: promise in setTimeout 2"));
}, 0);
 
Promise.resolve()
  .then(() => console.log("E: promise 1"))          // -> microtask queue
  .then(() => console.log("F: promise 2"));         // -> microtask queue (later)
 
queueMicrotask(() => {                              // -> microtask queue
  console.log("G: queueMicrotask");
  queueMicrotask(() => console.log("H: nested queueMicrotask"));
});
 
console.log("I: script end");                       // 2. sync
 
// EXECUTION ORDER:
// A: script start          (sync - call stack)
// I: script end            (sync - call stack)
// --- call stack empty, drain microtasks ---
// E: promise 1             (microtask)
// G: queueMicrotask        (microtask)
// F: promise 2             (microtask - chained from E)
// H: nested queueMicrotask (microtask - queued during G)
// --- all microtasks drained, pick macrotask ---
// B: setTimeout 1          (macrotask)
// --- drain microtasks (none) ---
// C: setTimeout 2          (macrotask)
// --- drain microtasks ---
// D: promise in setTimeout 2 (microtask from C)

Practical Patterns and Pitfalls

javascriptjavascript
// PATTERN 1: Yielding to the event loop
// Use setTimeout(fn, 0) to let pending I/O and rendering happen
 
async function processLargeArray(items) {
  const results = [];
 
  for (let i = 0; i < items.length; i++) {
    results.push(heavyComputation(items[i]));
 
    // Yield every 100 items via macrotask
    if (i % 100 === 99) {
      await new Promise((resolve) => setTimeout(resolve, 0));
      // This creates a macrotask, allowing:
      // - Pending I/O callbacks to fire
      // - Browser to render a frame
      // - Other macrotasks to run
    }
  }
  return results;
}
 
function heavyComputation(item) {
  let result = 0;
  for (let i = 0; i < 10000; i++) result += Math.sqrt(item * i);
  return result;
}
 
// PITFALL 1: await does NOT yield to macrotasks
async function awaitPitfall() {
  console.log("A: start");
 
  // await creates a MICROTASK, not a macrotask
  await Promise.resolve();
 
  console.log("B: after await");
  // B runs as a microtask, BEFORE any setTimeout callbacks
}
 
setTimeout(() => console.log("C: timeout"), 0);
awaitPitfall();
// Output: A: start -> B: after await -> C: timeout
 
// PITFALL 2: Promise.resolve() vs new Promise()
// Promise.resolve(value) schedules .then as microtask immediately
// new Promise((resolve) => resolve(value)) does too, but the
// executor runs synchronously
 
console.log("1");
new Promise((resolve) => {
  console.log("2"); // This is SYNCHRONOUS (executor runs immediately)
  resolve();
}).then(() => console.log("3")); // Microtask
console.log("4");
// Output: 1, 2, 4, 3
 
// PITFALL 3: Thenable objects add extra microtask ticks
const thenable = {
  then(resolve) {
    resolve(42);
  }
};
 
Promise.resolve(thenable).then((val) => console.log("thenable:", val));
Promise.resolve(100).then((val) => console.log("direct:", val));
 
// Output: direct: 100 -> thenable: 42
// Thenables add an extra microtask tick for unwrapping
 
// PATTERN 2: Immediate vs deferred execution
function scheduleWork(callback) {
  // Runs BEFORE I/O callbacks (microtask)
  queueMicrotask(callback);
}
 
function deferWork(callback) {
  // Runs AFTER I/O callbacks (macrotask)
  setTimeout(callback, 0);
}
 
// PATTERN 3: Batching DOM updates with microtasks
class DOMBatcher {
  #pendingUpdates = [];
  #scheduled = false;
 
  update(element, property, value) {
    this.#pendingUpdates.push({ element, property, value });
 
    if (!this.#scheduled) {
      this.#scheduled = true;
      // Batch all updates and apply in one microtask
      queueMicrotask(() => this.#flush());
    }
  }
 
  #flush() {
    for (const { element, property, value } of this.#pendingUpdates) {
      element.style[property] = value;
    }
    this.#pendingUpdates = [];
    this.#scheduled = false;
  }
}
FeatureCall StackTask Queue (Macrotask)Microtask Queue
StructureLIFO stackFIFO queueFIFO queue
ProcessingRun to completionOne per loop iterationAll drained per cycle
PriorityImmediateLowHigh (after call stack)
SourcesFunction callssetTimeout, I/O, eventsPromise.then, queueMicrotask
Can starveBlocks everythingCan be starved by microtasksCan starve macrotasks
OverflowRangeError (stack)Memory limit (unbounded)Memory limit (unbounded)
Rune AI

Rune AI

Key Insights

  • The call stack runs synchronous code to completion before anything else can execute: No callbacks, microtasks, or macrotasks can interrupt a running call stack frame
  • The microtask queue is drained completely after each macrotask and after the initial script execution: Promise chains resolve fully before timers or I/O callbacks fire
  • Only one macrotask is processed per event loop iteration, followed by a full microtask drain: This ensures microtasks always get priority over pending macrotasks
  • Recursive microtasks can starve macrotasks and block rendering because the queue must empty before proceeding: Use setTimeout to yield to the event loop during heavy microtask work
  • Understanding the priority hierarchy (call stack, nextTick, microtasks, macrotasks) is essential for debugging async timing bugs: Most unexpected execution ordering stems from confusing microtask and macrotask scheduling
RunePowered by Rune AI

Frequently Asked Questions

Why are microtasks processed before the next macrotask?

This design ensures predictable promise chain resolution. When you chain `.then()` calls, each continuation should complete before unrelated callbacks (like I/O or timers) run. Without this priority, promise chains could be interleaved with setTimeout callbacks, making async code unpredictable. The specification requires that all microtasks are drained between macrotasks to guarantee that promise-based code behaves consistently.

Can microtasks block rendering in the browser?

Yes. The browser only renders between macrotasks, after all microtasks have been drained. If microtasks keep scheduling more microtasks (recursively), the browser cannot paint a frame until they all complete. This is why you should use `setTimeout` or `requestAnimationFrame` to yield to the rendering pipeline during heavy work, rather than relying on promise chains or `queueMicrotask` alone.

What is the maximum call stack depth in JavaScript?

It varies by engine and platform. V8 (Chrome/Node.js) typically allows around 10,000-15,000 frames depending on the function's local variable count. Each frame consumes stack memory proportional to its local variables and closures. SpiderMonkey (Firefox) has similar limits. Exceeding the limit throws a `RangeError: Maximum call stack size exceeded`. Tail call optimization (TCO) can eliminate stack growth for tail-recursive functions, but only Safari implements TCO in production.

How does process.nextTick differ from queueMicrotask in Node.js?

`process.nextTick` callbacks run before microtasks (Promises). This means `process.nextTick` has even higher priority than `Promise.then`. The processing order in Node.js is: drain all nextTick callbacks, then drain all microtasks, then move to the next event loop phase. Recursive `process.nextTick` calls can starve both microtasks and macrotasks. The Node.js documentation recommends using `queueMicrotask` for most cases and reserving `process.nextTick` for specific low-level needs.

Conclusion

The call stack, task queue, and microtask queue form the foundation of JavaScript's concurrency model. Synchronous code runs on the call stack until it completes. Microtasks (promises) are processed with high priority after each call stack frame and between each macrotask. Macrotasks (timers, I/O) run one at a time with microtask checkpoints between them. For the event loop that orchestrates these queues, see JavaScript Event Loop Internals Full Guide. For how libuv provides the underlying I/O mechanism, explore Understanding libuv and JS Asynchronous I/O.