Summary
The user is new to Flyte and is seeking guidance on configuring permissions for managing resource usage. They want to establish a system where workflows and tasks that require high CPU or GPU resources are paused and can only be executed with approval from FlyteAdmin. Conversely, they want users to be able to run workflows that require moderate resources independently using pyflyte. The user inquires about setting an "Approval Policy" for resource requests that exceed a certain threshold and asks if such a policy can be implemented in Union or the open-source version of Flyte. They also mention that in Union, resource requests beyond the available resources will automatically fail, as it is aware of the deployed cluster.