History, Settings and Edit
History (H)
H opens the history of the focused building, or every incident in the fleet.
- Incidents: each period a target was degraded (over a warning threshold) or offline, with its reason, start and length, recorded as it happened.
- Charts: CPU, memory, fullest disk and traffic over 24 hours, 3, 7 or 30 days. Gaps mean "no data", never zero.
To see the whole city at one of those moments, use time travel.
Settings
Ctrl+K → Settings…
| Setting | Default | Meaning |
|---|---|---|
| Keep history for | 7 days (1–90) | metrics and incidents; older rows are deleted |
| Browser notifications | off | alerts by Web Push, tab closed |
| Alert channels | none | alerts: email, Slack, Telegram… |
| CPU warning | 90 % for 3 polls | a server turns amber |
| Memory warning | 90 % | |
| Disk warning | 90 % (any disk) | |
| Poll every | 2 000 ms | how often each server is measured |
| Stale after | 3 polls | missed polls before data freezes |
Thresholds are written to hosts.yaml (comments kept), so the file stays the source of truth.
Edit (E)
Focus a server and press E: name, district, tags, and its own warning thresholds (a database server that always uses 95 % memory need not be amber). The id never changes.
Resources: name, district, tags, and new credentials. Removing a resource deletes its saved credentials too.
Audit log
Ctrl+K → Open audit log. Append-only (the database refuses edits and deletions): connections, host key refusals, terminals opened and closed, servers and resources added or changed, AI commands run or denied, runbook runs, Doctor fixes applied, alerts sent. Never secrets.