Update: We will delay the implementation of v5 rate limiting until further notice
Effective February 28, 2020, the deprecated Kraken v5 API will be rate limited to an IP address basis of 80 requests per minute. In other words, requests from each IP address using your application are rate limited to 80 requests per minute. These new rate limits will not apply to existing Extensions that were submitted for review on or before January 31, 2020 until June 31, 2020. We may further reduce rate limits as we continue to migrate from the deprecated API to the New Twitch API and will post further updates on the forum if that is the case.
We are aware of many applications that may be affected based on total activity and are reaching out to several of these developers proactively. 80 requests per minute per IP address should cover most use cases and allow for scalability. Of course, always open to feedback and further understanding of all use cases here for the product teams.
What about authenticated requests per user? Do those have a 80/min bucket each or do we have to migrate everything to an incomplete helix API within 2 weeks?
The migration from the V5 API to the new one is a good thing, however in the current state there is a lot missing in the new API compared to the old one.
For the most part, these are essential things that are impossible to achieve with the new API.
There is also the fact that it is necessary with the new API to make several calls to obtain information obtained via a single call on the old (V5) API.
That’s a separate issue, not all fields from v5 are intended to be included in Helix, some fields have been intentionally not included so you should plan your apps accordingly for some fields no longer being accessible when v5 is removed.
what do you mean plan your apps accordingly? how do you expect me to plan to something that i will simply not have after? WHAT$ do you expect me to pull creation date out of my ass or do i have to open user card page which may or may not load for every user manually, make a program that captures my screen, scans the date and make an api based on that? i bet this fucking shit would have a better rate limit than 80/1m HOW THE hell do you plan to something that is GONE
WHAT? TRIHARD____KFC42069 has followed the channel and posting suspicious messages? OH LET ME JUST BRING MY MAGIC BALL SPIT AND BADOOGADONGO SLOP SLOP to see if this user just literally created the account to post a large illustration of N word in text form AKA THE NUMBER 9 or is he a 7 year veteran with 457 subs to various streamers. oh wait… do i have to actually wait when he does it? surely all these people are using the web chat which crashes if a slight wind of gust even tries to look at its direction and dont see removed messages and even if they are… THE MESSAGE IS STILL VISIBLE FOR A FEW MOMENTS. what is even the point then?
Put me down as a vote toward providing feature parity before nudging (shoving?) everyone off of V5. A lot of us have just been sitting here patiently - for years now - waiting for a viable upgrade path or any sort of concrete update as to what to expect from helix and when.
I’m sure they thought about all of this, and on the TwitchDev weekly they know and are aware of concerns, but they have said a few times now the V5 changes that are happening are because of changed to the back end that power V5, and V5 was originally made to power their website, and was opened to devs to use too. Helix is designed primary for 3rd party devs and engineered differently to not run into the issues v5 is having. They need to fix these to make sure there aren’t performance issues. They did say, they analyzed the logs, and identified the few devs that would be above the request limits on the dev weekly, and are trying to reach out to them.
I completely get it, and this isn’t going to affect my projects nearly as much as others, but we’re getting 20 days notice here. There are some features that Helix all-together doesn’t support and others like subscription totals and etc that have become significantly harder to scale. There’s been no hard shut-off date for V5 communicated (right?), theres been no official and definitive “Hey, we’re never going to support this again,” it seems - to me - like Helix has been treated as sort of as a WIP and we’re all kind of wondering where an when it will officially land - and how we’ll need to adjust.