Warning: This build has drifted.
The coverage report for this pull request build may be inaccurate because its base commit is no longer the HEAD of its target branch.
This means it includes changes from outside the original pull request, including, potentially, unrelated coverage changes.

    • Learn more: For more information on this, see Tracking coverage changes for pull request builds.
    • Fix now: For a quick fix, rebase this PR at GitHub. Your next report should be accurate.
    • Prevent going forward: To avoid this issue with future PRs, see these Recommended CI Configurations.

docs: Filter incompatible instance types; condense RFC

Switches the design from applying bandwidth weighting opportunistically
across a mixed NodePool to filtering instance types that don't support the
requested value, per review discussion.

Filtering removes most of the previous complexity: no capability-based
launch template grouping, no per-group omission of NetworkPerformanceOptions,
and no dependence on how EC2 treats the parameter on an unsupported type.
Karpenter never sends it to a type that doesn't accept it. The check follows
the existing CompatibleCheck pattern in pkg/providers/instancetype/compatibility
alongside the network interface and nested virtualization checks, and runs
during instance type resolution so no NodeClaim is created for a type that
would reject the request.

Mixed-generation fleets are now expressed with two NodePools and a weight
preferring the capable one, which is explicit in the spec rather than implied
by instance type capability.

NetworkInfo.BandwidthWeightings is the only source of truth for support, so
newer families (9th-gen already supports this) need no Karpenter release and
no static family list.

Also addresses review feedback on form: drops the table of contents, the
appendix, meta-commentary about earlier revisions, launch template hashing
internals, the duplicate NodeClass validation controller check, and the
repeated drift subsections. 445 -> 192 lines.

Adds an AI usage disclosure per Kubernetes contributor guidance.