How does load balancing work with multiple replicas

[pam2014]

(check apply)

• read the contribution guideline

Problem

i am trying to understand how does the round robin load balancing across elastic search data nodes work if the number of fluentd replicas > 1. Is the load balancing only for the data processed by a specific fluentd pod? Is there a recommended configuration on how to scale out aggregator pods while still able to achieve load balancing across multiple elastic search nodes

Steps to replicate

N/A

Expected Behavior or What you need to ask

Using Fluentd and ES plugin versions

Fluentd on k8s
Bitnami chart 5.0.15