When Envoy is disconnected from management server, we get these logs (mainly coming from statically defined EDS clusters or pending CLA requests)
[warning][config] bazel-out/k8-opt/bin/source/common/config/_virtual_includes/grpc_mux_subscription_lib/common/config/grpc_mux_subscription_impl.h:67] gRPC update for type.googleapis.com/envoy.api.v2.ClusterLoadAssignment failed

Since this is just a failed update (not reject), should we reduce this to debug? Any way, we do have stats like update_failure to detect this .

On a side note, should this reconnection to management server implement some kind of back-off strategy rather than a fixed aggressive interval?