Add graceful-restart support

[fedepaol]

Is your feature request related to a problem?

Whenever a version of metallb is updated, the speakers are rolled one at the time. This is causing the behaviour described here https://metallb.universe.tf/concepts/bgp/#limitations to be amplified for each node of the cluster.

Graceful restart should mitigate this problem.

Describe the solution you'd like

Either a per - peer graceful restart option, or a global one. To be discussed.
We should validate that all the proper signals are sent to the frr container and thus to the bgpd daemon (and e2e this), and cover the same feature in frr-k8s too.

Additional context

No response

I've read and agree with the following

• I've checked all open and closed issues and my request is not there.
• I've checked all open and closed pull requests and my request is not there.

[karampok]

Hi @fedepaol I can work on that. Thanks