Does AWS ElastiCache Redis Serverless support Celery as a Redis broker? Workers connect but never consume tasks.

I’m attempting to use Amazon ElastiCache Serverless for Redis (Redis 7.1, cluster‑mode enabled) as the message broker for Celery 5.3+ running on Amazon ECS Fargate. Celery powers asynchronous background tasks and periodic tasks (via Celery Beat) in a Django application.

The Problem

Celery workers:

• Successfully connect to the Redis Serverless endpoint
• Register themselves
• Create the expected queue keys (verified via redis-cli) …but they never consume tasks.

Symptoms:

• Tasks—including periodic Beat tasks—accumulate in the list (LLEN enterprise grows from 0→ 90).
• Running BRPOP celery 5 manually via redis-cli returns instantly, confirming basic connectivity.
• As a result, all asynchronous processing is effectively broken.

Environment / Configuration

• Python packages: celery>=5.3,<6 kombu>=5.3,<6 redis>=5.0,<6
• Celery is configured to use redis.cluster.RedisCluster.
• Because ElastiCache Serverless disables commands like CLIENT LIST and triggers CROSSSLOT errors, I apply a custom redis_monkeypatch.py.

Behavior:

• Without the patch → workers crash on startup with CROSSSLOT.
• With the patch → workers start normally, but the consumption loop never activates.

Core Questions

1. Is Celery + ElastiCache Serverless an officially supported or known‑working combination? I cannot find any AWS documentation confirming whether Celery works with the Serverless (cluster‑mode) flavor of Redis.
2. If it is possible, what is the correct configuration?

• Are specific broker_transport_options required?
• Are there Celery/Kombu settings needed to make RedisCluster compatible with Serverless restrictions?
• Are there known limitations with pub/sub or blocking list operations in Serverless mode?

Any guidance, shared experience or workarounds from others using Celery with ElastiCache Serverless would be extremely helpful.