DramatiqRabbitMQKubernetes

Broker connection pool exhaustion prevents scaling beyond 100 pods

critical
Connection ManagementUpdated Nov 21, 2025(via Exa)
Sources
Dramatiq Python Actors: Middleware Retries Throttling 2026johal.in
Technologies:
Dramatiqsubject
RabbitMQThe root cause of this issue originates in RabbitMQ
KubernetesKubernetes metrics correlate with this issue and help confirm diagnosis
How to detect:

Default broker connection pool exhaustion occurs when scaling Dramatiq workers beyond 100 pods. Each worker requires broker connections, and default limits prevent horizontal scaling at cloud scale.

Recommended action:

Configure ConnectionLimitMiddleware with max_connections=50 per worker. For larger deployments, implement RabbitMQ queue federation plugin to shard queues across clusters. Use broker.add_middleware(ConnectionLimitMiddleware(max_connections=10)) pattern.