IndicesClusterStateService currently manages its own retries when managing shard state. Channel related failures will now be managed in ShardStateAction (#15748, #15895) so the retry logic should be refactored there to only handle non-channel failures.
IndicesClusterStateService currently manages its own retries when managing shard state. Channel related failures will now be managed in ShardStateAction (#15748, #15895) so the retry logic should be refactored there to only handle non-channel failures.