Resolve name consistency of Redis instances/nodes

Today there's some confusion of how different dedicated Redis instances are called https://docs.gitlab.com/omnibus/settings/redis.html#running-with-multiple-redis-instances:

At the codebase we setup them as Gitlab::Redis::Cache (known by devs and infra as cache), Gitlab::Redis::Queues (mostly known by Sidekiq) and Gitlab::Redis::SharedState (also called persistence) by infrastructure.

These same names are now also used into log fields:

• redis_cache_duration_s
• redis_cache_read_bytes
• redis_cache_write_bytes
• redis_queues_calls
• redis_queues_duration_s
• redis_queues_read_bytes
• redis_queues_write_bytes
• redis_shared_state_calls
• redis_shared_state_duration_s
• redis_shared_state_read_bytes
• redis_shared_state_write_bytes

As well as yml files.

Would that be confusing enough to come to an agreement between dev/infra and make a considerable refactoring towards a consistent naming?