What problem does it solve?
A newly installed categraf agent is running, but the host never shows up in the Nightingale host list, or appears with unknown OS/CPU/version fields. This Skill pinpoints which segment of the 5-stage onboarding pipeline (local agent, heartbeat HTTP, server receive, target persistence, Redis/Prometheus metrics) is broken, instead of guessing at a single cause.
Core Features & Use Cases
- One-shot pipeline probing: Calls probe_target_onboard_status to gather evidence across all 5 onboarding segments in a single step, returning a pre-aggregated likely_segment and likely_causes.
- Segment-specific fix guidance: Decision table maps each stuck segment to concrete checks, such as heartbeat.enable, omit_hostname, TLS/BasicAuth errors, edge Redis configuration, and categraf version requirements (v0.2.35+).
- Ident label verification: Runs three PromQL variant queries (exact match, fuzzy match, instance-label fallback) to distinguish a true no-data condition from an ident label problem.
- Use Case: You deployed 3 categraf collectors via Helm but only 1 appears in Nightingale. The Skill probes the missing hosts, finds the heartbeat never reached the server, and gives you paste-ready journalctl and config.toml commands to fix it.
Quick Start
Ask the assistant why your newly installed categraf host does not appear in the Nightingale host list and provide the host's ident or name.
