javsalgar
commented
1 week ago Hi!
Thank you for using Bitnami. If I read correctly, it seems that you want the RabbitMQ prometheus plugin to expose a new metric, right? If that's the case, that would be something to check with the upstream RabbitMQ devs, as this is not related with the Bitnami packaging of RabbitMQ.
Name and Version
bitnami-rabbitmq-14.3.1
What is the problem this feature will solve?
Background:
I am using a Prometheus custom metric to Horizonal Pod Autoscale based on RabbitMQ Queue Depth, using the metric "rabbitmq_queue_messages_ready"
That's a fine metric, and I can use that, but I want to employ an algorithm that will correlate that to messages already consumed for that queue, thus we'd end up with:
1.) The average time it takes that queue to complete a task 2.) How many tasks we have left
We can then use that to see if the worker will fall behind and (read: plot the slope trend) and have it preemptively scale up rather than retroactively.
What is the feature you are proposing to solve the problem?
Make available a new metric that allows insight into messages consumed:
What alternatives have you considered?
I can just use "rabbitmq_queue_messages_ready" but it isn't as robust as correlating that with messages consumed.