[GEP-20] High Availability for reversed VPN connection

[timuthy]

How to categorize this issue?

/area control-plane
/area high-availability
/area networking
/kind enhancement

What would you like to be added:
Shoots which define spec.controlPlane.highAvailability.failureTolerance: {node, zone} get an HA control-plane but today the control-planes only have a single VPN connection.

The following points where sketched out after a discussion with @ScheererJ:

• VPN connections should be redundant: If one path is broken because of an outage, the other path must still work.
• Lifecycle management: Updates and configuration changes should be deployed in a rolling-fashion, so that there is at least one working tunnel at all times.

Why is this needed:
In case of a node or zone outage, a redundant VPN tunnel is necessary so that webhooks residing inside the shoot still work and can respond in a timely manner.

cc @ScheererJ @MartinWeindel

[timuthy]

/assign @MartinWeindel
Thanks for your support 🙂

[timuthy]

Completed by #6978
/close

[gardener-prow]

@timuthy: Closing this issue.

In response to this:

Completed by #6978
/close

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.