feat: NPD prober-like readiness plugin

[ajaysundark]

Implement a Node Problem Detector (NPD) custom plugin that runs a probe (similar to kubelet probes) and returns exit code 0/1 to indicate readiness. This allows operators who are familiar with pod readiness/liveness probes to reuse that configuration style for node-level readiness reporting.

• Reuses a familiar probe model so users can express readiness checks the same way they already do for pods.
• Can be deployed as an NPD plugin (lightweight go-lang binary to give exit status as per NPD
  requirements).
• This can be integrated into node-readiness-controller example
  (static-pod initialization for proper resource accounting #125) in the future.

Proposal:

• Add documentation and an example NPD plugin that:
  • Accepts probe configuration (HTTP/TCP like semantics) similar to kubelet probes.
  • Runs on the node and exits with code 0 for healthy, 1 for unhealthy.
  • Emits basic logs.
• Provide example config / NPD plugin manifest to demonstrate usage.

Acceptance criteria:

• A new example npd plugin implementation in examples/ or cmd/ that demonstrates an HTTP probe and returns exit codes.
• Doc updates for showing how to configure the plugin and how it maps to node conditions or consumption by node-readiness-controller.
• Tests to validate the plugin behavior.

[ajaysundark]

cc @GGh41th

[ajaysundark]

To clarify, the custom plugin design in itself is an exec model, so it doesnt make sense to include the "exec probe" here.

[santoshkal]

Hi,
Can I take this up? This would help me explore NPD and NRC.

/assign