Hello,
In my Loki deployments, I have three write pods, three read pods, and three backend pods, with the backend containing the compactor. I’m wondering if this is wrong in Loki? From what I’m seeing in the 3.7.1 docs, the compactor is supposed to be run as a single instance, is that correct? What happens if I have multiple instances? It seems like based on the memberlist, Loki itself designates one instances as the compactor. In that case, should compactor_grpc_address point to that one instance? How would know which of the three instances is designated as the compactor?
Thanks