Discovery Performance Issues on large clab topologies

[thediveo]

Some users with more demanding containerlab configurations run into discovery timeouts (30s). Apart from twisting timeout knobs, there is on one hand a potential performance problem with discovering port forwarding from large nftable rulesets, on another hand a needlessly detailed discovery when finding capture targets.

Thus, this issue tackles performance improvement of discovery in the specific context of capture targets.

• for the /mobyshark API endpoint use trimmed-down discovery, instead of the default exhaustive discovery.
• provide configurable (internal) discovery using discovery options in the same vein as lxkns does.
• provide scan configuration to NewNetworkNamespaces
  • opting out (skipping) port forwarding discovery