Monitor Linux with an agent that reports to one console.
Ubuntu, Debian, CentOS, Proxmox and friends. systemd unit, journal and failed-unit checks, ZFS pools on Proxmox. BE AI investigates when something fails; a person approves the fix.
install
One line, from the company's install page.
Run in a root shell. It fetches the signed binary, verifies the checksum, installs the service and connects. Run it again any time; it never overwrites the machine's identity.
curl -fsSL https://hub.[yourdomain]/install/inst_…/linux.sh | sudo bash
Outbound 443 to the hub only. No inbound rule, no VPN, no SNMP community.
what the agent watches on Linux
systemdFailed units listed, restart proposed by BE AI, approved by you.
JournalError-level lines on demand, capped in size.
DiskEvery mounted volume; on Proxmox the ZFS pool, because datasets all report the pool's free space.
ProcessesTop by CPU, or find one by name.
NetworkInterfaces, gateway, DNS servers, external address.
ChecksPing, DNS, URL, port, service, process, file age, free space.
questions
Before you put it on Linux.
Which distributions?
Anything with systemd on x86-64 or ARM64: Ubuntu 18.04+, Debian 10+, CentOS and Rocky 8+, Proxmox 7+. [Confirm before publishing.]
What does it need?
Nothing beyond the binary. No Python, no agent framework. About 8 MB on disk.
Put BE Hub on ten machines this week.
A pilot takes an afternoon: one install link, your own approvals, your own data. Uninstall the agent and you are out.