Either the limit was implemented hastily by higher ups or someone just messed up because the experienced developers were fired.
DDoS attacks are anything that can cause the service to not respond… In the malicious case, a lot of computers are made to try to access a service increasing the load to when the server can’t handle any more requests. Either keeping the server busy or make it shutdown.
Technically the limit doesn’t caused overall service outage, but for individual users.
The limit what they put in place didn’t considered that twitter itself is gonna call home automatically.
When you load the twitter page, you get a website that runs in your browser. That is a separate program from twitter servers. That program can communicate with the home server and ask for data, tweets pictures etc. That program can make a lot of requests in the background, like loading tweets in the background while you scroll so don’t have to wait for each tweet to load.
This a common practice to load a bunch of small data instead of one big chunk.
These background requests were counted as normal tweet requests by the limit monitor system. Essentially the twitter program itself made too much requests and locked itself out, because it hit the limit.
Poor execution that normal website calls and background API calls were not diffrerentiated.
Latest Answers