EAS Build Failure due to CDN (jsdelivr) Rate Limiting

Hi,

We’re facing an issue where EAS is failing the build due to a pod install failure.
The pod install failure seems to be caused by a TOO MANY REQUESTS response from the CDN (jsdelivr) it downloads from.

I tried to reproduce it by running pod install locally but weren’t able to, so it seems likely that your servers are being rate limited by the CDN?

Is there anything we could do to resolve this? We’re currently building a CICD pipeline for our new project using EAS, and failures like this could change our decisions in whether to adopt EAS due to maintenance cost like this, so hopefully this is not something that would occur frequently?

Also, for these kind of issues, is it okay to submit them directly to the eas-cli github repo or it’s preferred to be submitted here?

Thanks,
Leo

Hi

We plan to have some way of caching cocoapods, but I’m not sure when we will work on that. Can you link to job that failed for that reason?

Is there anything we could do to resolve this? We’re currently building a CICD pipeline for our new project using EAS

if it’s rate limiter on CDN side then not really, maybe retry on CI. It’s sth that we will need to look into on our side

so hopefully this is not something that would occur frequently?

I’m not sure how frequent this issue is, this is the first report we have seen with that problem.

Also, for these kind of issues, is it okay to submit them directly to the eas-cli github repo or it’s preferred to be submitted here?

It does not matter. On forum, if a question is in the correct category you have a higher chance that someone that works on that will see the post

2 Likes

Hi, thanks for the prompt response.
Here’s the screenshot as well as the 2 build ids that failed with this issue:
f5bd0dbe-6080-4849-970a-d69a337ae067
ee608a81-41aa-4dd1-a3f3-ed7ad91dfc04

We plan to have some way of caching cocoapods, but I’m not sure when we will work on that.

I understand that there’s uncertainty about when this will be picked up. But do you have a very rough guess of when it might happen? That sounds like something that will be very useful to address the rate limiting issue. Would it be weeks or months or years?

That sounds like something that will be very useful to address the rate limiting issue. Would it be weeks or months or years?

If it’s the rate-limiting issue, then it will be top priority to fix that because it affects everyone, but based on my attempts to reproduce that I’m suspecting that it might not be a case(I run about 20 builds, some of them concurrently without any issues).

The error you faced might have been a temporary issue with cdn itself. I know that you tried pod install locally, but it’s not really a valid comparison because you most likely had those pods already in cocoapods cache.

Either way, I’ll add some monitoring around this next week, currently, we have just 2 failed jobs that run one after another, it’s too small of a sample size to make any decision on.