I remember my tech lead pitching an idea about ML helping with detecting an anomaly - high load and accordingly scaling instances, and then scaling down later. Something similar can be done for detecting any kind of anomaly I think. CPUs, disks, memory
-
-
-
In principle, yes, but in practice nobody currently has anomaly detection that's anywhere near as good as a human looking at a graph and saying "That's weird..."
- Još 1 odgovor
Novi razgovor -
-
-
It's not about ML, people will take much longer to tolerate cars that won't exceed 40mph in a 40mph speed limit. I put my car in self driving mode and take my hands off the steering wheel and their anger is palpable and they don't understand why Im going so slow.
-
You can teach a car to drive... Good luck with the people...
Kraj razgovora
Novi razgovor -
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.