The Problem

Have you ever thought about what happens when your app tries to talk to a server, but the server is too busy or temporarily down? You might think to try again right away. But what if the server is still busy?

Trying over and over quickly can make things worse, using up resources and causing more delays. This is bad for both the server and the app.

Enter Exponential Backoff

Exponential backoff is a smart way to handle retries. It slowly increases the wait time between tries, which lessens the pressure on the server and raises the chance of making a successful connection. This approach finds a good middle ground between trying again quickly and not overloading the server.

When to Use Exponential Backoff

Exponential backoff is great for situations where systems might be temporarily down or too busy. Two cases where it can help:

Web apps talking to APIs.
Distributed systems where multiple different components are communicating.

Implementing Exponential Backoff in JavaScript

Let's take a look at a simplified JavaScript pseudo-code example to show how exponential backoff can be implemented:

function exponentialBackoff(attempt, maxAttempts) {
  if (attempt > maxAttempts) {
    console.log("Max attempts reached. Giving up.");
    return;
  }

  // Delay time in milliseconds
  // Increases as number of attempts increase
  // Resulting in higher delay for every attempt
  const delay = Math.pow(2, attempt) * 100;
  console.log(`Attempt ${attempt}: Retrying in ${delay}ms`);

  setTimeout(() => {
    // Your retry logic here, e.g., making an API call
    console.log("Performing retry...");

    // Assuming a function that checks whether the retry was successful
    const success = performRetry();

    if (!success) {
      // attempt increased here
      exponentialBackoff(attempt + 1, maxAttempts);
    } else {
      console.log("Success!");
    }
  }, delay);
}

// Example usage:
exponentialBackoff(1, 5); // Start with attempt 1, and a maximum of 5 attempts

Signs you may need exponential backoff

You might need to use exponential backoff if you notice a lot of network requests or API calls failing often, especially when it seems like the server or service you're using is too busy or has too much traffic.

Here are some specific signs:

Repeated Timeouts: If your requests often timeout, it might mean the server is too busy. Using exponential backoff gives the server time to recover and handle your requests better.
Server Errors: Getting errors like HTTP 500 (Internal Server Error) or 503 (Service Unavailable) means the server is overloaded or under maintenance. Exponential backoff helps by spacing out your retry attempts, easing the load on the server.
Rate Limiting Responses: If you're getting told you're sending too many requests (like with HTTP 429 Too Many Requests), you need to slow down. Exponential backoff helps you stick to these limits and manage retries more effectively.
Unstable Network Conditions: When the network is unreliable, like on mobile networks, exponential backoff can help by making sure retries don't make things worse by adding more load.

I want to clarify that you will not likely discover these signs during development. That's where logging and monitoring comes in. Being able to retrieve information and see how your system behaves in production.

What is Exponential Backoff?

Comments

More from this blog

Become a cracked product engineer today

Bipedal, humanoid, and the words for creature shapes in games

Lerp and smoothstep, what they actually do

IK and FK, what they actually are

How to make your textures fast on the GPU

The Problem

Enter Exponential Backoff

When to Use Exponential Backoff

Implementing Exponential Backoff in JavaScript

Signs you may need exponential backoff

Command Palette

Comments

More from this blog

The Problem

Enter Exponential Backoff

When to Use Exponential Backoff

Implementing Exponential Backoff in JavaScript

Signs you may need exponential backoff