Please share with the community what you think needs improvement with VMware RabbitMQ.
What are its weaknesses? What would you like to see changed in a future version?
One of the issues is that as soon as you go outside of a switch or not in IP address range, the clustering no longer has all the wonderful features so clustering outside of network boundaries is a problem. I'd like to see stream processing as an additional feature. Kafka has a streaming API and I'd like Rabbit to have that too.
RabbitMQ provides the ability to scale queues in a very simple and elegant way. If it had a “failure queue” with robust delivery and recovery built-in with the same power, that would be great. We use a completely different queuing system for failures. So there is a little more effort to take messages in a failure queue, analyze them, figure out what went wrong and then restart them in Rabbit. It is doable, and we do it, but if we had a round trip solution in Rabbit, that would be awesome. For me, having a robust failure queue, is high on the list of improvements needed in the near future. This is an important update needed because right now we are using Doctrine for our failure queue. Doctrine does a great job.
When you have complex tasks, RabbitMQ is hard to use. There are several things that you have to do manually, so there should be better tools for that.
Their implementation is quite tricky. It's not that easy to implement RabbitMQ as a cluster. It would be great if they could improve that.
I was struggling with installing a few things. It would be good if was somewhat similar to RedHat. There should be more documentation regarding installation troubleshooting. It's pretty straightforward, the setup, but it would be useful to know what to do if you do face certain challenges. Right now, without more in-depth documentation, it's unclear.
* Difficult to integrate with automated test and CICD * Moving beyond basic configurations can be challenging * Not clear how to implement durable subscriber connections * Not clear how a Rabbit service restart allows subscriber auto re-connect * Service cluster failover depends on shared disk infrastructure.
RabbitMQ is clearly better supported on Linux than it is on Windows. There are idiosyncrasies in the Windows version that are not there on Linux. The documentation for the Windows version is also less plentiful and less accurate. The online community clearly provides better Linux support, but this naturally follows from the smaller Windows installed base. There are also some potential concerns about how we maintain high-availability whilst also scaling out.
Which is better and why?
Let the community know what you think. Share your opinions now!