feat(gobgp): add kube_router_bgp_peer_info metric

[damex]

Replace the misleading kube_router_controller_bgp_peers gauge
which only counts 'cluster nodes'. with a new per peer metric
kube_router_bgp_peer_info with 'GaugeVec' that exposes actual
BGP session state from gobgp. labels include peer address, asn,
and session state. metric value is 1 if established and 0 otherwise.

Closes: #848

[damex]

made isolated testing for the changes on the same node

made simplified bird2 bgp config

protocol bgp kubernetes {
  local 10.128.76.16 as 64567;
  neighbor 10.128.76.16 port 180 as 64567;
  ipv4 {
    import all;
    export all;
  };
}

this is the metric before connection is established:

$ curl 10.128.76.16:9109/metrics |& grep bgp_peer
# HELP kube_router_bgp_peer_info BGP peer information
# TYPE kube_router_bgp_peer_info gauge
kube_router_bgp_peer_info{asn="64567",peer="10.128.76.16",state="active"} 0

and this is after:

$ curl 10.128.76.16:9109/metrics |& grep bgp_peer
# HELP kube_router_bgp_peer_info BGP peer information
# TYPE kube_router_bgp_peer_info gauge
kube_router_bgp_peer_info{asn="64567",peer="10.128.76.16",state="established"} 1

[damex]

i did see that workflows got an approval and received message about this:

https://github.com/cloudnativelabs/kube-router/actions/runs/22016039123/job/63619708028 pipeline here complained about not handling SESSION_STATE_UNSPECIFIED.

i originally map all others states and had fallback to 'unknown' (unspecified is supposed to be 'unknown'). so i have added mapping for this 'unspecified' case too and it satisfies linter.

i did run lint locally and it passed

damex@foosha:~/Documents/kube-router$ git status
On branch feature/implement-bgp-peers-info-metric
Your branch is up to date with 'origin/feature/implement-bgp-peers-info-metric'.

nothing to commit, working tree clean
damex@foosha:~/Documents/kube-router$ golangci-lint run
0 issues.

please run pipeline again

[damex]

maybe we want to distinguish here if it is ibgp or ebgp peer (internal/external) ?
something like

kube_router_bgp_peer_info{asn="64567",peer="10.128.76.16",state="established", type="ibgp"} 1

i can add it here or as a separate pull request if that's acceptable

[damex]

i thought about it and added type = ibgp/ebgp.
also renamed peer label to address since that seem to make much more sense here.

that's before establishing connection with bird2

$ curl 10.128.76.16:9109/metrics |& grep bgp_peer
# HELP kube_router_bgp_peer_info BGP peer information
# TYPE kube_router_bgp_peer_info gauge
kube_router_bgp_peer_info{address="10.128.76.16",asn="64567",state="active",type="ibgp"} 0

and that's after

$ curl 10.128.76.16:9109/metrics |& grep bgp_peer
# HELP kube_router_bgp_peer_info BGP peer information
# TYPE kube_router_bgp_peer_info gauge
kube_router_bgp_peer_info{address="10.128.76.16",asn="64567",state="established",type="ibgp"} 1

and this is ebgp with other 'host'

that's before establishing connection with bird2

$ curl 10.128.76.16:9109/metrics |& grep bgp_peer
# HELP kube_router_bgp_peer_info BGP peer information
# TYPE kube_router_bgp_peer_info gauge
kube_router_bgp_peer_info{address="10.128.76.15",asn="64568",state="active",type="ebgp"} 0

that's after

$ curl 10.128.76.16:9109/metrics |& grep bgp_peer
# HELP kube_router_bgp_peer_info BGP peer information
# TYPE kube_router_bgp_peer_info gauge
kube_router_bgp_peer_info{address="10.128.76.15",asn="64568",state="established",type="ebgp"} 1

[aauren]

Hey @damex - Thanks for the contribution, I've been slammed the past couple of weeks, but I'll try to take a look at this one and your other one for SCTP support as soon as possible.

[catherinetcai]

Thank you so much for your detailed notes and thorough testing @damex! I only have very, very minor comments here.

[aauren]

Hey @damex apologies for getting to this late.

Thanks for the contribution! Everything looks good here, and I agree with all of @catherinetcai's suggestions. I tried deploying it and it looks good to me.