Server monitoring
from one small shell script.
Keep an eye on CPU, memory, root disk, load average and uptime on the Linux or BSD hosts behind your sites. A readable POSIX /bin/sh script posts metrics over HTTPS every minute with a per-server token; no daemon, root access or binary agent is needed. Tested with simulated and local agent reports, not yet on customer hosts.
- From Solo
- Plain POSIX /bin/sh + curl
- Threshold and silent-host alerts
How it works
A push check for the host itself, alongside the external checks on your website.
Add a server in the app
Name the host, optionally link it to a site or client, set CPU, memory and disk warning percentages and choose how many silent seconds count as a missed report.
Schedule one cron line
Copy the one-time bearer token (stored only as a SHA-256 hash) and run the plain
/bin/shscript from cron or a systemd timer once a minute as an unprivileged user.Get warned when limits or reports fail
When CPU, memory or disk reaches its threshold, or when reports stop arriving past your silence window, Beamfar alerts your routed channels and sends a recovery notice once metrics return to normal.
What you get
Auditable POSIX shell script
Reads
/proc/stat,/proc/meminfo,/proc/loadavg,/proc/uptimeanddf -Pk /(with BSD and macOS fallbacks) and posts JSON withcurl.Hashed per-server tokens
Each server gets its own secret token shown once at creation or rotation; only its SHA-256 hash and short prefix are stored in PostgreSQL.
Threshold, silent and recovery alerts
Separate plain-language alerts when a resource crosses your warning limit, when the host goes silent and when it reports normally again, respecting active maintenance windows.
Client and site scoping
Link a server to a site or client so alerts follow client channel routing and read-only portal contacts see only their own client’s servers.
Which plans include it
Server monitoring is included on every paid plan, from Solo up.
- FreeNot included
- SoloIncluded
- TeamIncluded
- AgencyIncluded
- Agency PlusIncluded
Questions about server monitoring
Does the script need root privileges or background daemons?
No. It is a short POSIX /bin/sh script that reads standard system stats and sends one HTTPS POST with curl. Run it from an ordinary user’s crontab or systemd timer.
What does the script collect?
CPU usage percentage, memory usage percentage and total megabytes, root filesystem usage percentage and total gigabytes, 1/5/15-minute load averages, uptime in seconds and optional hostname. It never reads application files, environment variables or logs.
What happens during a scheduled maintenance window?
Metrics continue to be recorded in history, while threshold, silent-host and recovery alerts are held back for the organisation, client or linked site under maintenance.
Can I rotate or revoke a server token?
Yes. Rotating a token immediately invalidates the old hash and displays a replacement token once; pausing or removing the server rejects any further metric posts.
Watch the host alongside the website.
Create a server token, add one cron line and see CPU, memory and disk within a minute.
Included from Solo.