helm: Add support for HorizontalPodAutoscaller on router

[Lunik]

Description

This PR add new options in the Helm chart values allowing to deploy and HorizontalPodAutoscaler for the router component.

This is basically the implementation of this part of the doc : https://fission.io/docs/installation/advanced-setup/#create-hpa-for-router-service

By default, nothing change for the user.

User can select the Kubernetes autoscaling apiVersion allowing support for user with cluster version under v1.23. By default the autoscalingApiVersion used is autoscaling/v2 (support v1.23+ and required in v1.26+).
The router.autoscaling.metrics handle the full metrics structure regardless of the apiVersion.

I have placed the value autoscalingApiVersion at the root allowing a reuse if we want to add autoscaling to other components of Fission in the futur.

Which issue(s) this PR fixes:

Didn't find any

Testing

• Deploying the plateform without setting any values => same as before
• Deploying the plateform with autoscaling.enabled=true

kubectl get hpa

NAME     REFERENCE           TARGETS         MINPODS   MAXPODS   REPLICAS   AGE
router   Deployment/router   <unknown>/80%   1         10        2          9m14s

Checklist:

• I ran tests as well as code linting locally to verify my changes.
• I have done manual verification of my changes, changes working as expected.
• I have added new tests to cover my changes.
• My changes follow contributing guidelines of Fission.
• I have signed all of my commits.