Add RabbitMQ publishRate trigger

[rwkarg]

Signed-off-by: Ryan Karg [email protected]

Added an additional trigger for RabbitMQ scaling on publishRate.

This allows for higher throughput workloads to scale based on the incoming message rate. This is useful because the message backlog does not often correlate with the desired instance count.

Both publishRate and queueLength triggers can be specified on a single ScaledObject. I would imagine the normal case would be to use both of these triggers.

Checklist

• Commits are signed with Developer Certificate of Origin (DCO - learn more)
• Tests have been added
  [ ] A PR is opened to update our Helm chart (repo) (if applicable, ie. when deployment manifests are modified)
• A PR is opened to update the documentation on (repo) (if applicable)
  • document new publishRate trigger keda-docs#388
• Changelog has been updated

Fixes #1643

[rwkarg]

Comparison of two different but similar workloads over 24 hrs. The orange line on the graph correlates with the instance count. Both workloads are in the 1k-3k/sec. range and have smooth rates of change over the day so the instance count would expect to be smooth as well to match that.

Using only queueLength it bounces bewteen min. instances and a too-high instance count after a backlog builds up:

Using publishRate (as well as queueLength) it stays steady:

In the second example, we actually had an impact on a dependency that cause a backlog to build up so the queueLength trigger eventually resulted in a higher instance count that you can see in the two spikes at the beginning of the graph.

[ahmelsayed]

LGTM, thanks @rwkarg

[zroubalik]

LGTM, thanks @rwkarg !

[zroubalik]

Let's merge this once docs PR is finalized.

[tomkerkhove]

Thanks a ton @rwkarg!