Keepalive timeout for GRPC

Hi all,

I was trying to find a way in Kong GRPC setup for the GRPC client to detect GRPC server connection going down and vice versa, but could not find any viable way apart from having Upstream Passive Health checks to do some checking at least for Kong to Upstream connection! Is there a way to solve it somehow? Nginx can be configured to set up TCP keepalive for Nginx proxy to a client and Nginx proxy to server connections.

On the other hand, Kong proxy termination itself is well detected by GRPC client and server.
I appreciate any help/direction!