I am new to Kubernetes and currently deploy an application on AWS EKS.
I want to configure a Service
of my K8s cluster deployed on AWS EKS.
Here is the description of my issue: I did an experiment. I spin up 2 Pods running the same web application and expose them using a service (the Service is using LoadBalancer
type). Then I got an external IP of that Service. Then and found that requests that I sent were not distributed evenly to every Pod
under the Service
I created. To be more precise, I sent 3 requests and all the three requests are processed by the same Pod.
Therefore, I want to configure the load balancer algorithm to be round robin
or least_connection
to resolve this issue.
I asked a similar question before and I am suggested to try the IPVS mode of the kube-proxy
, but I did not get the detailed instruction on how to apply that config, and did not find any useful material online. If the IPVS mode is a feasible solution to this issue, please provide some detailed instructions.
Thanks!