Need help on ETCD Request Retry Mechanism

[rajassub]

I am trying to use ETCD client version client/v3 v3.5.6.
ETCD cluster has 3 members A,B,C.

Need to know how load balancer supports retries on failure for PUT/GET/DELETE requests.

assume client request is sent to A by balancer, if A returns a error or down, how many times balancer will retry the request to same A, to next member B before returning failure to client.

Once a failure is returned to client, if client retries the same request, will balancer send the request again to A, or now send it to B