Document the concurrency model for subscribers

[00yhj22-debug]

Description

Closes #2749.

Adds a dedicated docs page (docs/docs/en/getting-started/subscription/concurrency.md) that explains the two independent levers FastStream offers for concurrent processing — max_workers and RabbitMQ's prefetch_count — and clears up the specific confusion called out in the issue:

Does prefetch_count=10 mean ten messages are processed at the same time?
→ No. It means the broker may deliver up to ten before requiring acks; how many are processed in parallel is determined by max_workers.

The page covers:

• What each setting does on its own.
• How they combine (with a small table of the four common combinations).
• A note that Kafka/Confluent have no prefetch_count and link to the existing "Concurrent processing" notes on the Kafka subscriber docs.
• A note that Redis/NATS/MQTT only have max_workers.

Also wires the new page into docs/docs/navigation_template.txt after the existing "Dynamic Subscribers" entry.

Type of change

• Documentation (typos, code examples, or any documentation updates)

Checklist

• My code adheres to the style guidelines of this project
• I have conducted a self-review of my own code
• I have made the necessary changes to the documentation (this is the docs change)
• My changes do not generate any new warnings
• I have added tests to validate the effectiveness of my fix
• Both new and existing unit tests pass successfully on my local environment

No tests added — the change is a new docs page plus a single-line nav entry.