Today in CAA, we have the ability to select the best-fit instance from among a set of specified instances based on cpu and memory.
This issue is to track the support for using GPUs in instance selection based on annotations. This functionality is dependent on Kata sending the required annotations to the CreateVM call and is tracked here - https://github.com/kata-containers/kata-containers/pull/10453
Today in CAA, we have the ability to select the best-fit instance from among a set of specified instances based on cpu and memory. This issue is to track the support for using GPUs in instance selection based on annotations. This functionality is dependent on Kata sending the required annotations to the CreateVM call and is tracked here - https://github.com/kata-containers/kata-containers/pull/10453