Raj-glitch-max/kubernetes-llm-incident-response-benchmark
Benchmark measuring whether an LLM actually reads cluster evidence before diagnosing a Kubernetes incident, or just guesses confidently. Chaos injection, evidence ablation, cross-model scoring.
GitHub repository with 120 stars and 1 forks.
Language: Python
Topics: benchmark, chaos-engineering, incident-response, kubernetes, llm-evaluation, observability, sre