Lol with kubernetes et al style container orchestration + service architecture, I’d say this is almost becoming more common. It’s just so easy to automatically recycle a pod if one of the processes starts being too greedy.

Yeah maybe don’t use LLMs