Container ActiveGate not sharing load

[gopher]

Hi All, 

We have noticed that the elected leader of the container Active Gate is not sharing load with other Active Gate replicas.
This is causing the Leader to get overloaded from a resource point. 

We have increased replicas from 2 -> 4 and still only 1 (the elected leader) appears to be doing all of the agent routing and Prometheus scraping.   

Is this known or noticed by anyone else?  does anyone have a solution other than trying to split functionality into single large AG container instances using different deployments in different namespaces. 

I'd like to avoid having to run them similar to our VM based AGs. 

[ChadTurner]

We live by noncontainerized AG's. We found that we have more power and control over it. 

[gopher]

HI @ChadTurner,  we have both, but it defeats the purpose of having a Kubernetes native solution and keeping all traffic (except AG egress) local to the cluster. 
Building an AG based on legacy approach, has significant overhead around managing network / firewall rules / NACL, build, additional hosts (cost), architecture, security ...  and everything else that goes along with it if I have to build outside of the cluster - a perfect example of #Dynatoil. 

At the moment, I am moving towards having OTEL collectors which decouples me from Dynatrace and gives me more flexibility with the solutions that we can interact with, shifting out Prometheus / Kubernetes metrics and Trace and only leaving agent routing and inbuilt Kubernetes monitoring on the AG.  Not exactly what I want, but benefits are quickly adding up.