Healthcheck

Use this skill for evidence-first operational diagnosis.

A useful healthcheck does not just say healthy or unhealthy. It identifies scope, collects live evidence, separates symptoms from causes, and leaves the operator with a prioritized fix-and-verify plan.

Gather the minimum context first

Identify:

• system in scope: host, VM, container, service, app, network path, or full stack
• environment and criticality: prod, staging, dev; customer-facing or internal; uptime sensitivity
• reported symptom or decision needed: diagnose outage, assess risk, verify deployment health, review hardening, or spot bottlenecks
• recent changes: deploys, config edits, dependency updates, infra events, cert renewals, incidents
• evidence available: service status, logs, metrics, traces, dashboards, config, health endpoints, backup state
• resource posture: CPU, memory, disk, file descriptors, network, queue depth, latency, error rate, saturation
• protection posture: auth exposure, secrets handling, TLS, firewalling, backups, restore confidence, observability coverage
• constraints: maintenance windows, rollback options, access limits, and blast radius concerns

If evidence is missing, say the conclusion is provisional and name what would resolve it.