Error handling in JavaScript is a topic that evokes strong opinions, and I’m here to share mine: the traditional try-catch approach is clunky, inconvenient, and outdated. At Garmingo, where we built Garmingo Status — a SaaS solution for uptime and infrastructure monitoring—we’ve shifted away from try-catch blocks. Instead, we embraced a TypeScript-based approach that provides predictable, standardized responses for asynchronous operations.
This article shares why we believe this paradigm is a game-changer for developer productivity and how it helped simplify our codebase. While it’s an opinionated take, I hope it inspires you to rethink how you handle errors in your own projects.
Let’s face it: error handling in JavaScript can get messy. Traditional try-catch blocks come with a host of challenges:
Here’s a simple example highlighting these issues:
try { const user = await fetchUser(); try { const account = await fetchAccount(user.id); console.log(account); } catch (accountError) { console.error('Error fetching account:', accountError); } } catch (userError) { console.error('Error fetching user:', userError); }
The result? Code that’s harder to read, debug, and maintain.
At Garmingo Status, we ditched try-catch in favor of a standardized response structure for all asynchronous operations. Here’s how it works:
Every async function returns a Promise with a predefined union type:
Promise< | { success: false; error: string } | { success: true; result: T } >;
This approach ensures that:
Here’s the same example from above, rewritten with this pattern:
const userResponse = await fetchUser(); if (!userResponse.success) { console.error('Error fetching user:', userResponse.error); return; } const accountResponse = await fetchAccount(userResponse.result.id); if (!accountResponse.success) { console.error('Error fetching account:', accountResponse.error); return; } console.log(accountResponse.result);
As you can see it does not introduce any nesting for the main logic of your app. It just adds these small checks for error handling, but the main flow remains uninterrupted and can continue like there was no need for error handling in the first place.
The biggest benefit is knowing exactly what to expect. Whether the operation succeeds or fails, the structure is consistent. This eliminates the ambiguity that often comes with error objects.
Gone are the days of deeply nested try-catch blocks. With the typed approach, you can handle errors inline without breaking the flow of your code.
The structured approach makes your code cleaner and easier to follow. Each operation clearly defines what happens in success and failure scenarios.
TypeScript’s static analysis ensures you never forget to handle errors. If you accidentally omit a check for success, the TypeScript compiler will flag it.
No approach is without its drawbacks. The typed response paradigm requires you to explicitly check the success status for every operation, even if you’re confident it will succeed. This adds minor overhead compared to the traditional approach, where you might simply avoid error handling altogether (albeit at your own risk).
However, this “drawback” is also one of its strengths: it forces you to think critically about potential failures, resulting in more robust code.
At Garmingo, this approach has transformed how we build asynchronous utilities and libraries. Every API call and database query adheres to this standardized response structure, ensuring consistency across our codebase.
In fact, EVERY single async function that is reused trough-out the project and could fail uses this approach.
The result? A smoother (and much faster) development experience and fewer late-night debugging sessions.
For example, a fetch function could look like this:
try { const user = await fetchUser(); try { const account = await fetchAccount(user.id); console.log(account); } catch (accountError) { console.error('Error fetching account:', accountError); } } catch (userError) { console.error('Error fetching user:', userError); }
This predictability has been a game-changer for our team, allowing us to focus on building features rather than untangling error-handling logic.
Traditional try-catch blocks have their place, but for modern JavaScript development — especially in TypeScript-heavy codebases — they’re often more trouble than they’re worth. By adopting a typed response paradigm, you gain predictability, readability, and peace of mind.
At Garmingo, we’ve seen firsthand how this approach simplifies development and enhances our ability to deliver a polished product like Garmingo Status. While it might not be for everyone, it’s an approach I strongly believe more developers should consider.
So, are you ready to rethink error handling? Let me know your thoughts!
The above is the detailed content of You're Doing Error-Handling Wrong!. For more information, please follow other related articles on the PHP Chinese website!