The nightly pass.
Once a day, at a time you choose per company, BE AI reviews every included machine that is online.
Reads a day of history first
Heartbeats, events, check results, the last pass. Cheap, instant, and enough to know what changed.
Then looks at the machine
Diagnostics, services, recent errors, checks, network facts. Read-only.
Speaks only when it must
Healthy machine: nothing to say, nothing charged beyond the read. Something drifting: a short note, and a proposal if a command would fix it.
every heartbeatWhat is watched without asking.
Disks with a date
Growth per day from the last week and a day the disk will be full. On the dashboard, on every fleet row, and in the alert when it comes.
Memory and swap creep
Trends per machine over 7 and 30 days, so a leak shows as a slope, not a crash.
Services and reboots
A service that stops, a reboot that was not planned, a machine that comes back on an old release.
Custom checks every minute
Mapped drives, printers, VPN peers, file age, free space, listening ports. Drift is caught before a person notices.
Behind on releases
Machines not on the latest signed agent, and a roll-out to fix it by tag or company.
Degraded storage
A ZFS pool that goes from ONLINE to DEGRADED is a critical alert before the second disk fails.
on request or on a scheduleRecipes a technician would run if there were time.
Patch health
Pending reboot, failed updates, last successful update, uptime. Ask for it on a machine or a whole company.
Service health
Stopped auto-start services that should be running, crash events, restarts proposed only where a service clearly belongs up.
Health summary
"Give me a health summary of this machine": patching, disk, memory, services, and anything from the event log worth a look.
Verify the effect
After any approved change, one click asks BE AI to check the machine again and confirm what changed.
questions