GQL request handler redesign #652
Labels
Enhancement
New feature or request
Comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Hello,
This is a self-note issue tracker for me, to remind myself to rework the GQL request handling code, so that composite requests are each handled separately. This is due to the fact that the requests can be made in batches up to 20 (even though I've seen the API handle up to 27 requests just fine, although that's probably only due to some kind of a response size limit), and if only one of them fails, all 20 are repeated again on a retry. This is not only really unnecessary, but also puts more strain on the Twitch servers for no real reason.
A proper implementation should including creating a request handler delegate for each request, then batching them together (most likely via timed polling, AKA "wait for X time before sending the request, just to see if the code adds more delegates in the mean time"), and then error handling and retry being separate from the operations that have finished successfully.
The text was updated successfully, but these errors were encountered: