It would be amazing for sysdump analysis to get a generated infrastructure architecture overview/network diagram (graphviz?).
Is your feature request related to a problem?
Not directly, but it often already takes quite some time to gain a good overview of infrastructure from a sysdump before one is actually able to start digging into finding the actual Cilium issue.
Describe the solution you'd like
Something like this would make it easier and faster to understand the infrastructure context before digging into the actual Cilium issue. In my imagination, a generated network diagram could include information like this:
Kubernetes Nodes
Name
Node IP
L2 of the Nodes
Region / AZ (if present)
Perhaps the node gateway IP as an additional "router" node in the L2?
Cilium / Envoy Pod per Node (so it's directly clear which Cilium Pod runs on which Node)
Cilium Internal IP
Cilium Health IP
If present:
Cilium BGP Peers with IPs
Cilium Ingress / GatewayAPI LB Service IP
Etc.
This diagram could either be generated during the cilium sysdump sub-command or after a sysdump has already been taken, and we use a separate sub-command to achieve this.
Proposal / RFE
It would be amazing for sysdump analysis to get a generated infrastructure architecture overview/network diagram (graphviz?).
Is your feature request related to a problem? Not directly, but it often already takes quite some time to gain a good overview of infrastructure from a sysdump before one is actually able to start digging into finding the actual Cilium issue.
Describe the solution you'd like Something like this would make it easier and faster to understand the infrastructure context before digging into the actual Cilium issue. In my imagination, a generated network diagram could include information like this:
This diagram could either be generated during the
cilium sysdump
sub-command or after a sysdump has already been taken, and we use a separate sub-command to achieve this.