server-doctor

Host auditing for Telegram automation builds an evidence-based environment map by collecting service identities, runtime owners, and access methods before deeper troubleshooting. It inventories systemd, launchd, Docker/Compose, and direct process workloads across reachable and unreachable machines.

Diagnosing false outages on OpenClaw gateway hosts requires applying strict anti-false-outage evidence discipline. This approach structures diagnosis around spec correctness versus ops quality, using an evidence hierarchy to classify health and prevent incorrect outage declarations.

Runtime discovery detects how bot workloads are started and supervised across systemd, launchd, Docker/Compose, direct processes, and terminal multiplexers. It identifies the supervisor mechanism for each service to produce a complete user-by-user runtime inventory.

Incident triage on production servers requires a preflight intake establishing reachable targets. You must provide which machines run which services, the Unix owner account, the likely runtime style like systemd or Docker, and the safe connection method such as SSH or Tailscale.

Public-safe operational maps instruct redaction of sensitive data while producing host-by-host maps, bot-to-host mappings, and restart, log, and status commands. The documentation references incident and health classification rules without exposing system secrets.

Unreachable hosts during environment mapping are handled by running the audit in an unreachable mode. The process classifies access status, documents known unknowns, and records the target in the operational map without forcing false diagnostic conclusions.