Advise against min-masters

see
rabbitmq/rabbitmq-server#10627 (reply in thread)

README.md

@@ -5,7 +5,7 @@
 ### Branches
 
 The `main` branch is the production branch. Commits to it are deployed
-automatically to www.rabbitmq.com by a Cloulflare worker.
+automatically to www.rabbitmq.com by a Cloudflare worker.
 
 Older versions of the docs that we don’t want to host in Docusaurus to limit
 the number of versions are put in branches of the form `v3.13.x`, `v4.0.x`,

docs/mqtt.md

@@ -682,6 +682,9 @@ Also, routing messages with fewer topic levels is faster.
 1. In workloads with high subscription churn, increase Mnesia configuration parameter `dump_log_write_threshold` (e.g. `RABBITMQ_SERVER_ADDITIONAL_ERL_ARGS="-mnesia dump_log_write_threshold 20000"`)
 1. When connecting many clients, increase the maximum number of Erlang processes (e.g. `RABBITMQ_SERVER_ADDITIONAL_ERL_ARGS="+P 10000000`)
 and the maximum number of open ports (e.g. `ERL_MAX_PORTS=10000000`).
+1. Do **not** use [configuration](./configure) or [policy](./parameters) `queue_master_locator = min-masters`.
+`min-masters` reads all queue records from RabbitMQ's database to decide on which node to place a classic queue, which becomes expensive for clusters containing many queues.
+Note that `queue_master_locator = min-masters` is a default configuration as of rabbitmq/cluster-operator v2.7.0.
 
 Consult the [Networking](./networking) and [Configuration](./configure) guides for more information.