Hi all, I noticed in the scaling doc page (https://github.com/kserve/modelmesh-serving/blob/main/docs/production-use/scaling.md) that now is possible to set the `ServingRuntime `autoscaling with `HPA`, but using metrics based on cpu utilization. Is it possible to scale the `ServingRuntime` using metrics regarding GPU? Thanks in advance
Hi all,
I noticed in the scaling doc page (https://github.com/kserve/modelmesh-serving/blob/main/docs/production-use/scaling.md) that now is possible to set the
ServingRuntimeautoscaling withHPA, but using metrics based on cpu utilization.Is it possible to scale the
ServingRuntimeusing metrics regarding GPU?Thanks in advance