Hi Salmaan. Thanks for your post.

How about elasticity? When we scale up the number of replicas for example. Will the client schedule a call to the headless service to refresh its cache so that it can “discover” the new replica?