AAA retries changed silently



The thing I’m going to complain about is a shame. I know that being rude doesn’t help to resolve an issue. But here, it’s not a bug, it’s a recurring bad behavior. Twitter dev platform shows again its incapacity of fulfilling basic good practice. I hope that this topic will help you folks to improve yourselves; or at least the quality of the API/docs.

Webhook retries count changed recently, and silently.
It was enjoyable to have 7 retries in AAA webhooks, mostly because it was done among a 4 hours window; allowing few hours of downtime without missing data irremediably (archive: []).

New version: 3 retries among 5 minutes, you (co-devs) better not have a 5+ minutes downtime in your infrastructure, because it would cost you to lose data without being able to recover it.
BTW: you forgot to adapt the If you do not confirm validation within these eight attempts, sentence.

No trace of this change in the changelogs, that’s a pity.

About “Client must use other Twitter endpoints to recover data.”
It is currently impossible to retrieve any (a n y) like of a tweet. Many devs requested that feature among the past, the only way of doing it was to use the AAA. But if the p a i d p r e m i u m v e r s i o n of the AAA cannot propose good retries, what can we do in the case of a 10m downtime? Nothing. Data is lost, forever.

So, intelligent people working for Twitter, please:

  • be communicative
  • be empathic & good-willing
  • propose good & exhaustive third-party endpoints

The “get likes” endpoint exists for Twitter Official (Overly Premium) clients. It is (for example) the /statuses/:id/activity/summary.json endpoint, reserved to old apps, new ones probably now use the version 2 of the API. Some tutorials & gists out there encourage to use these endpoints with [illegitimately obtained] official tokens.

My question is: why don’t you do something to ease our life using your API? Is it because of third apps not displaying your ads? You already slayed them enough by shutting down the user stream.

Thanks for your sense of comprehension.


Hi Richi - thank you for your candid feedback. I agree that if we make any changes to our infrastructure, it should be communicated to our developers. I’ll sync with the relevant folks on this and try to come back with more information.


Hello @richie3366 - Despite what docs may have indicated in the past, no code changes have taken place related to the retries functionality since the launch of the product. Over the past month, we had noticed that the documentation did not reflect reality, so the change to the docs was to fix an inaccuracy.

We are aware of the limitations of the existing retry window and the challenge it poses to developers, so we are actively thinking about alternative ways to serve any missed data in the event of down period greater than 5 minutes. If you haven’t already, please check out our stay informed page, as we will be sending updates via these sources as we release new functionality.


Thanks @Hamza & @LeBraat for taking the time to give an answer.

Quite frankly, it is deplorable that inaccurate docs took place for so long, and that it had never been accurate on this point. I’m not your boss and I have no right to blame the individual(s) responsible for this discrepancy, but it adds to the long list of regrettable (not to say frustrating) things about the Twitter API service.

I can name many APIs deserving congratulations for their good management, accuracy & completeness, for example the Stripe API. If the webby awards had a Best API award, it would certainly go to them. Being the best is hard, yes, but you may try not to be the worst. :slight_smile:

I hope that you’ll propose solutions in the near future, I am already subscribed to the dev feeds. I’ll be happy to see the coming improvements. I hate complaining about things, but when the demonstrations of “not-giving-a-damn”* replies (!) &/or mistakes are chaining indefinitely, it’s getting hard to express it kindly.

(* : The very long list of topics asking for interesting features but answered for years by “It is not planned, take a look at our roadmap [topic closed]”)

Thanks again for your support!


I appreciate your candor and feedback. Our team does strive to do better work than this. We have several active work streams aimed at developing better principles around our docs products that will, in my opinion, be a dramatic improvement upon our existing offering. Until we roll out these new programs, please do let us know if you see any other confusing concepts in the docs. We will make sure to address any confusing resources so that you can continue developing amazing products.

Thanks again,
Dan and the Twitter dev team