How would you match traffic from the same source without keeping the record of that source?
That’s a technique not a goal. What are you trying to do?
Find when a specific endpoint / AS / country starts sending dos levels of traffic, (or hack attempts) so they can be banned.
Rate limiting prevents a specific IP from causing a successful DoS. You can log higher level information like county without linking it to a specific user.
In terms of hacking, building a secure site prevents this problem at the source. Banning specific IP’s in a world of proxies and public WiFi is almost useless.
You don't ban them forever. Banning specific ranges which impact you right now is very effective too. Also "building a secure site" at some scale is impossible. At some point you try to figure out where the risk is, how to mitigate it, and what happens after a break-in. You can't prevent it. Logging helps track specific behaviour and catch those situations. That's similar to fraud prevention as well. The fact that someone who just logged in from Germany tries to spend credit in a request from Brazil is important and prevents real crime. That kind of information needs to be connected to an account.
Which devolves your argument into collecting this information is significantly more profitable. Which I think is generally accepted as true, but not nessisarily enough to make it acceptable.