Added ScaledObject parameter StartReplicaCount to scale from 0 to N

[ldimarzi]

This pull request introduces a new parameter, startReplicaCount, to the ScaledObject in order to change the scaling behavior from 0. The goal is to facilitate scaling from 0 to N replicas, instead of the default 0 to 1.
This change can be useful in scenarios where applications need fast responsiveness when is activated.

Checklist

• I have verified that my change is according to the deprecations & breaking changes policy
• Changelog has been updated and is aligned with our changelog requirements
• A PR is opened to update our Helm chart
• A PR is opened to update the documentation on (repo)
• Commits are signed with Developer Certificate of Origin

Relates to #

PR Helm chart: kedacore/charts#606
PR Documentation: kedacore/keda-docs#1308

[JorTurFer]

Hello!
Thanks for the contribution, but this is already supported: https://keda.sh/docs/2.13/concepts/scaling-deployments/#idlereplicacount

You can set idleReplicaCount: 0 and then, KEDA will scale from 0 to minReplicaCount and from minReplicaCount to 0.

[ldimarzi]

Hello @JorTurFer ,

Appreciate your response. With startReplicaCount set to 5, this feature allows scaling from 0 to startReplicaCount. Subsequently, the scaling behavior from startReplicaCount to 0 is gradual, proceeding in example through the sequence 5 -> 4 -> 3 -> 2 -> 1.

Contrary to this, using idleReplicaCount permits scaling only from minReplicaCount to 0.

I'd appreciate your thoughts on these changes considering this description.

[JorTurFer]

I think that the best way to manage new features is opening an issue to discuss about it.
I see the benefits of this, but I'd like to know the thoughts from other folks.
@zroubalik @tomkerkhove

Given that and talking just about the code, it looks good but the changelog has to follow the style guide and test coverage is required (unit test, e2e test or both)

[zroubalik]

Yes, please start with an issue first, honestly I don't understand the feature. Imho this behavior is solved by setting for example by settingidleReplicaCount: 0 and minReplicaCount: 5. What am I missing here?

[ldimarzi]

Hello @zroubalik,

Thank you for your reply. When the startReplicaCount is configured as 5, it enables scaling between 0 and the specified startReplicaCount value. Then the scaling process from startReplicaCount to 0 can be decremental and gradual, in example following sequence such as 5 -> 4 -> 3 -> 2 -> 1 replicas.

In contrast, utilizing idleReplicaCount restricts scaling to occur solely between minReplicaCount:5 and idleReplicaCount:0 without decrementing gradually replicas.

I'd appreciate your thoughts