Currently, I have a frontend
, backend
, and tf-serving
service. My current configuration exposes both frontend
and backend
services outside.
The service tf-serving
, however, is an internal service only used by the backend
service. They communicate using the gRPC API provided by TensorFlow Serving.
Would it be possible for tf-serving
to be load balanced too such that when I call it from the backend
service, Traefik will load balance it for me?
BTW: I deploy these containers on a docker swarm with 10 VMs.