Comment by jklinger410 | Hacker News Reader

Comment by jklinger410 | Hacker News Reader

Parent Full thread

> Twitter is firing off about 10 requests a second to itself to try and fetch content that never arrives because Elon's latest genius innovation is to block people from being able to read Twitter without logging in.

Hey, just as an FYI, this comment doesn't make any sense.

Why would blocking people from viewing the TL on the home page cause Twitter more stress?

I think the GP is saying Twitter's client-side scripting doesn't understand permanent failures. It thinks all failures are transient, so it keeps retrying, hoping to get through. That creates more traffic than if Twitter just let the user read the site.

> Why would blocking people from viewing the TL on the home page cause Twitter more stress?

If you block stuff somewhere in the microservie hell in the backend, but forget to update the frontend code properly, and everything was built with the assumption that authentication and authorization is not a thing, and suddenly it is, there's bound to be craploads of stuff that can start spamming network requests.

I don’t think ‘built with the assumption that authentication and authorisation is not a thing’ is the best description. Requests from logged in users would already need to be authenticated to see private accounts they follow. It’s more just the client having aggressive retry logic in the case of whatever rate limit responses they’re getting.

> Hey, just as an FYI, this comment doesn't make any sense.

You need to understand that there is a huge gulf between "this doesn't make any sense" and "I don't understand this".

I don't know what all is involved on the backend, but when I go to twitter.com it runs 123 requests and transfers 11MB of data in order to serve me a login page.

Browser:can I get the tweet?

Server: No

Browser:can I get the tweet now?

Server: No

Browser:can I get the tweet now?

Server: No

Browser:can I get the tweet now?

Server: No

Browser:can I get the tweet now?

Server: No

Browser:can I get the tweet now?

Server: No

Browser:can I get the tweet now?

Server: No

Browser:can I get the tweet now?

Server: No

Browser:can I get the tweet now?

Server: No

You know what I mean?

10 If User has reached end of scroll region, load more content

20 Wait $timeout for new content; if a load failed, it was probably a mobile network glitch or something else we should ignore

30 GOTO 10

I think he was trying to say that the Frontend of the Twitter timeline is not aware of the new request limit, and when that limit is reached it just retries the request, as it would in a normal network error, and keeps in retrying resulting on DDOS.

But I don't believe on that because Twitter is a billion dollar company and that kind of mistake is dumb. So idk.

You severely overestimate musk’s software engineering prowess

> But I don't believe on that because Twitter is a billion dollar company and that kind of mistake is dumb.

This sort of mistake is very common especially when a product is going through a period of change, so it seems very believable to me that recent rate limiting features (or older features that have not been used in a long time or maybe ever) on the backend hadn't been adequately tested to make sure there are no unforeseen side-effects on the front-end when they are activated under pressure.

Very common for start ups with inexperienced devs, not for long standing tech companies with the best and most experienced employees...

Isn't Elon supposed to know this shit himself?

I guess we learned alot about the state of the company as well as elons own coding skills with this event, didn't we?

Even in “long standing tech companies with the best and most experienced employees” this sort of issue has been known to fall through the cracks. Twitter has experienced significant staffing changes, procedure changes, and code changes, recently and any one of those three things can exacerbate the chances of side-effects of changes not being realised until too late.

When you buy a company and your staff has had near complete turnover and your product significantly changes (at least in consumers eyes), are you actually the same company?

I'd say that shortly after Elon bought it twitter effectively became a start-up rather than an established company.

Arguing otherwise is like saying Xerox Park would have been the exact same institution even if all the staff were swapped with interns and management fundamentally change mission statements etc.....

Kind of absurd when framed that way, huh?

Then you didn't look at how any of the Atlassian products are built. Most of them are multiple implementations of the same Frontend-Franework on the same page because independent teams work on stuff. Then you have hundreds of request "just to load some issue list".

Probably twitter can even do worse.

I wouldn't have believed it but we're seeing it happen in real time haha. How twitter has fallen from the glory days

The people who had the knowledge to realize this was a dumb mistake were probably fired by Musk.

If it doesn't make any sense, what's your analysis? At least they tried to come up with some technical explanation and incite my curiosity, rather than dismiss it outright.

It could be retry logic that gets stuck in certain conditions that have been met by how TL blocking is handled.

Broken retry logic due to an unexpected situation.

Reply on news.ycombinator.com