zlacker

[return to "Twitter Is DDOSing Itself"]
1. Topfi+Ew[view] [source] 2023-07-01 21:09:47
>>ZacnyL+(OP)
Speaking from very painful, personal experience, few things are more agitating than being forced to execute on something you fully know is a horrible idea, especially when you tried and failed to communicate this fact to the individual pushing you to go against your best judgement.

Even more so when that person later loudly proclaims that they never made such a request, even when provided with written proof.

I can of course not say whether the people currently working at Twitter did warn that the recent measures could have such major side effects, but I would not be surprised in the slightest, considering their leadership's mode of operation.

Even as someone who very much detests what Twitter has become over the last few months and in fact did not like Twitter before the acquisition, partly due to short format making nuance impossible, but mostly for the effect Tweets easy embeddability had on reporting (3 Tweets from random people should not serve as the main basis for an article in my opinion), I must say, I feel very sorry for the people forced to work at that company under that management.

◧◩
2. martin+oN[view] [source] 2023-07-01 23:11:50
>>Topfi+Ew
The people at Twitter who understood the system and could predict the side effects were all fired or left. My guess is Elon said "the site's too slow!" Engineers noticed that the home feed request was slow. They didn't understand how it worked, had no tools to profile it, and were given an unrealistic deadline to fix it. So about the only thing they could do was issue multiple, parallel requests and hope that at least one of them was fast.

I worked in the games industry for a while, and came to understand how they could spend so much money and so much time, and yet release a game where even basic functionality was broken. It's exactly this sort of extreme schedule pressure that, ironically, makes a huge morass where changing one thing breaks 10 other things, so progress grinds to a halt.

◧◩◪
3. jayd16+o71[view] [source] 2023-07-02 02:21:41
>>martin+oN
> The people at Twitter who understood the system

But this is Scaling-101 stuff. It's not some super complex or unique system going wrong. At least according to the article, it's a classic case of bad retry logic leading to a death spiral.

https://en.wikipedia.org/wiki/Thundering_herd_problem

◧◩◪◨
4. PaulDa+Na1[view] [source] 2023-07-02 02:54:43
>>jayd16+o71
This has absolutely nothing to do with the thundering herd problem.
◧◩◪◨⬒
5. jayd16+vh1[view] [source] 2023-07-02 04:16:30
>>PaulDa+Na1
Explain why not, if you please. If unresponsiveness causes increased traffic, which causes further unresponsiveness, is that not referred to as a thundering herd problem? Is the stated mitigation of a backoff not fully relevant here?
◧◩◪◨⬒⬓
6. inepte+ek1[view] [source] 2023-07-02 04:52:17
>>jayd16+vh1
It's the difference between one customer asking a hundred cooks for a waffle and a hundred customers asking one cook for a waffle. The former is the thundering herd (a bunch of processes trying to do something that only needs to be done once, causing resource contention) and this is akin to the latter (with the "customers" being parallel requests from the frontend).
[go to top]