Hello, I'm in the process of transferring our image generation from a competitor to runpod.
My understanding is that the lifecycle of workers is that after they initialize they become idle until they receive a request. Upon receiving a request, they become both active and running.
This is the part I'm conflicted on:
In practice, I'm observing that once the request completes, the worker stays active, but not running, for as long as your idle timeout is set. Once the timeout is reached, the worker becomes idle.
However, screenshot attached states that "you are charged for active and idle workers at the same rate", and implies that "idle" is equivalent to being active but not running. I did some testing and this isn't true: when a worker is marked as "idle", there is no charge for it.
Is this a mistake in the option descriptions show in the screenshot? Am I misunderstanding completely?
Recent Announcements
Continue the conversation
Join the Discord to ask follow-up questions and connect with the community
R
Runpod
We're a community of enthusiasts, engineers, and enterprises, all sharing insights on AI, Machine Learning and GPUs!