Hey,Our use case is a cluster of 3 instances of RabbitMQ Broker that serve around 10000 connections over MQTT. The connections of each client are not stable.
Each client that connects uses the Auth BE plugin when it connect/publish/subscribe.
All MQTT messages are sent with QoS 1.
Our brokers are hosted in AWS.
Our LB strategy's is Round-Robin, since it is the only strategy available by ELB.
I'm trying to figure out what is the optimal AWS instance type to use in our use-case.
I'm considering three options:
C3.XLarge - With 4 Virtual CPUs, Memory 7.5 GB, SSD Storage of 40GB
R4.XLarge - With 4 Virtual CPUs, Memory 30.5 GB, Networking up to 10 Gigabit, EBS Storage linked to IOPS SSD storage.
M4.2XLarge - With 8 Virtual CPUs, Memory 32GB, EBS Dedicated Bandwidth 1000Mbps, EBS will be linked to IOPS SSD Storage
Has someone ever tackled such use case and can share his insights?
Does someone has suggestions of how to monitor IO done by RabbitMQ? We run it over CentOS7.
Thanks,
Sasson