What problem does it solve?
When a Kubernetes service gets no traffic or an Ingress stops routing, the cause can hide anywhere along the chain from Ingress to Service to selector to pods to EndpointSlice readiness. This Skill walks every link in that path so you get a complete, evidence-based diagnosis instead of a partial check that misleads the next engineer.
Core Features & Use Cases
- Full Path Tracing: Systematically checks Ingress, Service, selector-to-pod matching, EndpointSlices, and readiness, marking each link as checked or not checked.
- Differential Diagnosis: Distinguishes look-alike symptoms such as selector mismatches, unready pods, port mismatches, NetworkPolicy denials, and Ingress backends pointing at non-existent Services.
- Readiness Accounting: Reports ready and not-ready endpoint counts separately so a Service with zero serving capacity is never reported as healthy.
- Use Case: A user reports the
web service in namespace app1 is unreachable. The Skill retrieves the Service, matches its selector against pods, inspects EndpointSlices, and reports that 3 pods match but none are Ready β a readiness-probe problem, not a wiring problem.
Quick Start
Ask the agent to trace why the Kubernetes service named web in namespace app1 is receiving no traffic and report every link it checked.
