simulating node problems for Image testing

[sanjayk0508]

I am planning to perform manual Kubernetes testing on the node-problem-detector image (installed via Helm) to simulate node issues, similar to problem-maker. However, if I understand correctly, problem-maker cannot be used for image-based testing.

I wanted to check if there is an alternative approach to manually test the node-problem-detector image and validate its functionality.

For a comprehensive test, we aim to cover:

1. Simulating Node Problems – Ensuring NPD detects and reports issues such as OOM kills, Docker daemon hangs, etc.
2. Prometheus Metrics Testing – Verifying that NPD correctly exports metrics related to node problems for monitoring systems.

This is similar to e2e testing, but I haven’t found any clear workarounds specifically for testing the node-problem-detector image. Are there any recommended methods or existing approaches that I'm missing out? Sorry if similar issue has been raised before.
Any guidance or suggestions would be greatly appreciated!!