Different use case from most I assume:
I have multiple proxy alloy instances configured in a cluster behind a load balancer to support nodes without internet connectivity in a datacenter.
Should metrics sent to the cluster be sent to each member? Round-robin?
Documentation explains the use/benefits of clustering with regards to sharing scrape jobs, but not when it comes to prometheus.receive_http workloads.
If you are using receive (which i believe is somewhat equivalent to remote_write) then you don’t need your alloy agents to be clustered. And I’d imagine your standard load balancing method would work, whether it’s round-robin or less connection.