Skip to main content

Has anyone else noticed that that the rate limit headers sent with all API responses changed after the outage yesterday? Previously the headers were CamelCase (e.g. X-RateLimit-Limit) and now they are all lowercase (e.g. x-ratelimit-limit).

Did I miss an announcement or is this just a surprise change? And I guess if you're reading this, heads up that the headers appear to have changed so if you have any rate limiting logic be sure to update it.

I think the reason is:

HTTP/1.1 and HTTP/2 header names are case-insensitive but HTTP/2 enforces lowercase header name

 


Right you are. I guess something changed on their backend to start following the lowercase spec since they used to send camelcase headers.

Admittedly it's also an issue on my end that the language I'm using does case-sensitive string comparisons on the headers.


I think they simply disabled the http1.1 support in AWS Cloudfront. 
But yep, we had the same problem. Took a short moment to understand the reason.

Now I only hope their other API issues are being fixed soon. 


Reply