The rate limit headers aren't returned to users

[matt-aitken]

If there's a rate limit error, it would be good if the rate limit headers are added to the error object. Currently it's not possible to get them when using the SDK.

[emroot]

I second that. Right now error code 429 has not been been added to the typescript definition either. Would be great to be added.

[iBobik]

I my case it will be best if the SDK just wait a second, so it will process all emails finally.

[iBobik]

How about this solution?

• The Resend class will track requests on hold
• fetchRequest will check how many request are on hold and sleep for requests*500ms
• On 429 error it will retry (including waiting again)

This should work for one instance od the Resend class.

• Do we need to handle multiple clients like in horizontal scaling deployments or multiple clients?
• Are the limits recorded for IP address, API key or account level?

[gabrielmfern]

@iBobik

Are the limits recorded for IP address, API key or account level?

It's on team level, so the team has the rate limit for all API keys that can be increased. We also have this page for it https://resend.com/docs/api-reference/rate-limit

• The Resend class will track requests on hold
• fetchRequest will check how many request are on hold and sleep for requests*500ms
• On 429 error it will retry (including waiting again)

I think it's worth it more for us to just expose the existing headers in a clear way, instead of providing a batteries-included solution like that. I've made the pull request #562 that pushes in the direction I mentioned, would love to know what are the thoughts of everybody here about it.

Would basically look like this:

const response = await resend.emails.send(/* ... */);

//  limit: number;
//  remainingRequests: number;
//  shouldResetAfter: number;
if (response.rateLimiting?.remainingRequests === 0) {
	setTimeout(() => {
		// ...continue making requests...
	}, response.rateLimiting.shouldResetAfter * 1000);
}

And for a failed, rate limited, request:

const response = resend.emails.send(/* ... */);

if (response.error?.name === 'rate_limit_exceeded') {
	setTimeout(() => {
		/* ...code for retrying... */
	}, 1000 * response.error.retryAfter);
}

[gabrielmfern]

Leaving this open until we release it on a stable version

[gabrielmfern]

We've just released this onto a canary version that you all can try resend@6.1.0-canary.0. Here's how to use it to check the remaining requests on the rate limiting window, for example:

const response = resend.emails.send(/* ... */);

console.log(response.rateLimiting?.remainingRequests);

And here's how to get the right amount of seconds to retry a failed request, for example:

const response = resend.emails.send(/* ... */);

if (response.error?.name === 'rate_limit_exceeded') {
  console.log(response.error.retryAfter);
}

Just letting you know to get your feedback and thoughts on this, can't confirm this is going to be the API we're going to release with.

[DylanLyon]

Would love to see this implemented! I'm running into rate limit issues myself right now and I'd love to be able to properly implement it with the returned retryAfter header!

[gabrielmfern]

We've switched it up to just exposing the headers, and it's on canary now, resend@6.4.0-canary.1. Feel free to give some feedback here. It should be in a stable release around next week.