Complete operations map
The bot is an operator UI, not a bypass. Make one deliberate change, review its deploy plan, and confirm the resulting revision only after verifying the intended path.
| Section | Use it for | Check before leaving |
|---|---|---|
| Status | revision, convergence and pending work | revision and warnings are expected |
| Devices | issue, reissue, revoke and select egress | profile reaches only the intended operator |
| Tunnels | inspect, enable, disable and test | a probe succeeds before production traffic |
| Subscriptions | inspect candidates, accept one or restore known-good | health result and selected upstream |
| Routes | explain destination-to-tunnel decisions | each private destination has its intended path |
| Between devices | add or remove a narrow client ACL | source, target, protocol and port are intentional |
| Hub | hub, fallback and probe settings; key rotation | rotation is scheduled and profile impact is understood |
| Host and Logs | diagnose services and restart the agent | active revision and failure signal are recorded first |
Deploy and incident path
Section titled “Deploy and incident path”The Deploy screen shows current and proposed revisions plus a confirmation deadline. Confirm only after a probe works; otherwise let the timer restore the previous revision or choose rollback explicitly.
- Record the active revision in Status.
- Inspect the affected tunnel or subscription health result.
- Review remediation in Deploy.
- Confirm after a successful probe, or restore/rollback.
- Use Host and Logs when the agent itself is failing.
Notifications
Section titled “Notifications”Settings persists per-category alerts for automatic rollbacks/deploys, agent errors, convergence, tunnel health, host drift, subscriptions and out-of-band changes. Reduce noise by disabling only informational categories; keep critical failure alerts enabled.