Collect
Persist application, cloud, reliability, deployment and external evidence with provenance.
How OpsChugex diagnoses operational problems.
A diagnosis is useful when an operator can inspect the evidence, understand uncertainty and prove whether the system recovered.
Correlate
Connect observations in the Evidence Graph without labeling temporal association as causation.
Diagnose
Rank deterministic hypotheses and expose support, contradictions, gaps and alternatives.
Verify
Proof of Fix compares post-change evidence against baseline before verified recovery enters Operational Memory.
What OpsChugex refuses to pretend.
- Correlation is not proof of causation.
- A confidence score is not a mathematical probability.
- Missing evidence remains visible.
- Business impact is not invented when traffic, transactions or revenue are not measured.
- User feedback does not silently rewrite the original diagnosis.
- Automatic production remediation is disabled.
Measure before marketing.
OpsChugex separates ordinary operator feedback from controlled benchmark cases. It does not publish a global accuracy percentage from a handful of incidents.
Provider status is explicit.
AWSLIVE_READ_ONLY
KubernetesLIVE_READ_ONLY
AzureCOLLECTOR_READY
Google CloudCOLLECTOR_READY
Kubernetes is LIVE_READ_ONLY after Docker Desktop Kubernetes validation. No EKS cluster is currently attached. Azure and Google Cloud remain COLLECTOR_READY until their own real-environment validation.