Jorge Turrado Ferrero

Any update?

Any update on this? can we close the issue?

Hi @laithalissa This makes sense totally, are you willing to contribute adding the feature?

I think that it was done by error IDK when, but I agree that as they (length and rate) are separated options in the scaler, we should use the correct...

It should be `>` because if users don't provided any value for it means that the value is any message in the queue, so we set 0 as default. `isActive...

I agree with you and having only publish rate without queue length isn't a real world scenario, but IMO the user should be who decides the scaling behavior. With the...

I've checked current logic and I think that we should refactor this in future versions https://github.com/kedacore/keda/blob/19f14a82319b1b0aa7d524b1474e149a2231f117/pkg/scalers/rabbitmq_scaler.go#L618-L624 I think that if I have chosen a public_rate mode, the scaler has to...

I think that this issue is related with rounding and tolerance problems. Once you've reached 88 instances, you won't never reach 89 because the variation (1012m over 1000m) is 0,012%...

> A candidate for Troubleshooting/FAQ? Could be

Hello! > In addition, if you use ScaledObjects exclusively with CPU/memory triggers, the following metrics are not recorded at all: > - keda_scaler_active > - keda_scaler_errors > - keda_scaler_metrics_value >...