[Translation] [Crowdin] Use synchronous HTTP requests when reading the Crowdin API #48109
Conversation
|
If the limit is 20 concurrent requests and not 1, we don't need to use synchronous requests. We only need to limit concurrency to 20. |
|
Can you try forcing HTTP/1.1? Because the default is to open max 6 connections the remote hosts. If we hit the 20 limit, it means we're using HTTP/2. If that works, then the fix could be just to force the |
|
Thanks for your suggestions, putting a limit of 6 or 20 concurrent requests won't really work in our case, since we pull translations locally, in our CI, and when deploying our app in review apps / staging / production. Putting a limit of 1 would be ideal for us, this way we can be sure that we won't hit the Crowdin's concurrent requests limit across envs. |
|
Maybe the provider should also wrap the client into a RetryableHttpClient |
Can you still give it a try? 3x6 = 18 so that might work. |
|
Closing as we're likely not going with this approach. Please let us know if force http 1.1 is enough in practice. |
|
Yeah I gonna try to force HTTP 1.1, or playing with max host connections option ASAP |
Hi!
We are migrating to Crowdin and we faced
429 Too Many Requests issueswhen pull translations, because we have a lot of locales and domains in our application, and Crowdin limits 20 parallel requests per account.This is like #44416, but for Crowdin :)
Thanks.