docs: Improve BGP Control Plane page

[krouma]

Please ensure your pull request adheres to the following guidelines:

• For first time contributors, read Submitting a pull request
• All code is covered by unit and/or runtime tests where feasible.
• All commits contain a well written commit description including a title,
  description and a Fixes: #XXX line if the commit addresses a particular
  GitHub issue.
• If your commit description contains a Fixes: <commit-id> tag, then
  please add the commit author[s] as reviewer[s] to this issue.
• All commits are signed off. See the section Developer’s Certificate of Origin
• Provide a title or release-note blurb suitable for the release notes.
• Thanks for contributing!

Fixes: #23817

Clarifies installation of BGP Control Plane
Fixes non-functional hyperlinks

[qmonnet]

Thank you, the changes look all good.

Could you please fold the new commits (4 and 5) into the relevant commits (1 and 2 I think)? You can drop the Co-authored-by: tags, in particular because GitHub generates them with my personal email instead of work email.

[qmonnet]

Looks good to me, thank you!

Re-adding review request to Dylan, which was presumably removed by accident.

[krouma]

Yes, I just clicked to rerequest review and didn't notice it removed the second one.

[qmonnet]

Dylan: Just thinking, while reviewing this PR again: If the two different BGP features are mutually exclusive, should we consider a follow-up PR to make the Cilium agent fail to start if both are required by the user?

[dylandreimerink]

To be honest, I don't know if the two are mutually exclusive, I will double check this for myself. However, I can recall a few issues where I have seem users enable both at the same time and them being confused as to why configuration from the legacy feature doesn't apply to the BGP Control Plane.

I think there fundamentally isn't a conflict since both use different methods of configuration and do not interfere on a network level as far as I know. And I would not enforce mutual exclusion in the agent since that would make migrating from MetalLB to BGP CP harder.

I still agree with the spirit of this doc change, but I suggest we change it be advisory and keep the clarification that enabling BGP CP does not disable the legacy BGP feature.

[github-actions]

This pull request has been automatically marked as stale because it
has not had recent activity. It will be closed if no further activity
occurs. Thank you for your contributions.

[joestringer]

While two different BGP modes might not technically be exclusive, if we can't come up with a legitimate use case to combine the two, then it may simplify support and user expectations if we mark them as mutually exclusive anyway. If someone really has such a use case, they could raise the discussion point and we could consider relaxing the restriction at that time.

[joestringer]

@krouma is the feedback clear at this point? From a PR triager perspective, I see the following:

• sig-bgp still needs to approve the PR
• The full testsuite needs to be run
• Given that the last push was about 2 months ago, it's likely that the PR will need a rebase before we can successfully run the full CI with /test.

EDIT: After posting the comment, I realized this is a docs-only PR, so full CI does not need to run. This is just blocking on sig-bgp feedback.

@dylandreimerink does this improve overall codebase health? Are you happy with the proposal or do you see specific things that need to be changed to ensure that the proposal improves the health of the tree/docs?