👋 Not sure if this is the right channel or if i have to go into one of the specific tools channel. But, what is the recommended way for on-prem scaling of e.g. hydra or similar ? Would it be virtual scaling, so simply add more cpu/ram etc or horizontal scaling? I'm kinda thinking of a "simply" 3 node setup as a starting point with a HAproxy LB/Reverse proxy in front of it. And then a 3 node Postgres for Hydra in this case.

the docs don't specifically recommend either way, but ory is built for trivial horizontal scaling, so that's the approach we're taking with our setup - https://www.ory.sh/docs/ecosystem/software-architecture-philosophy

that is also easier i would say

hello there 🙂 We generally don’t provide “best practices” for scaling as it is mostly user-case based. Some benefit more from less instances with more power, other with a more of a micro-service setup. The apps support both types, but as they are rather low resource consuming, we would advice to scale the workloads horizontally 🙂

👍