Know which server is about to fail, and why, before anyone logs a call.

The SERVERMON agent runs on each Windows server as a LocalSystem service and collects its readings locally. Every request it makes is outbound and signed, so no inbound port, VPN, or stored domain credential is needed. The install is a single command, which means it can be rolled out through GPO, Intune, or whichever RMM tool you already use, and new servers wait for an admin to approve them before they appear.

SERVERMON reports free space for every volume in gigabytes and as a percentage, and projects how many days remain before it fills at the current rate. Alongside that it watches SMART failure prediction, file-system error events, blue screens, and unexpected shutdowns, so the warning signs of a dying disk or an unstable server are raised as alerts instead of sitting unread in an event log.

The dashboard groups servers the way you organise them, by site, role, or client, and shows each one's status, volume usage, CPU, memory, and network at a glance, updating live. Opening a server shows performance over the last hour up to the last 90 days, storage projections, the top processes by CPU and memory, recent events, hardware and software inventory, and its alert history.

Every rule has a warning level, a critical level, and a sustained period. Values cascade from sensible defaults to the organisation, then to a group, a server, and finally a single volume, with the most specific setting winning and a one-click reset back to the inherited value. A CPU spike that lasts thirty seconds does not page anyone; ten minutes at 98% does.

Each rule can send to WhatsApp, and every message says which server, what crossed the line, and when. Daily caps and per-recipient hourly caps put a ceiling on how many messages can go out, a usage meter shows the estimated cost in rand, and delivery status is tracked per recipient. When a whole site drops, SERVERMON recognises it as a possible site outage and says so once, instead of sending one alert per server.

An alert moves from open to acknowledged to resolved, and its timeline records each step: when it opened, when it escalated from warning to critical, who acknowledged it and with what note, and when it cleared. Alerts can be filtered and exported to Excel or CSV, and the audit log records every change to servers, thresholds, users, and recipients.

A server can be muted for anything from fifteen minutes to seven days while someone works on it, maintenance windows suppress alerts for planned work, and monitoring hours keep non-critical noise out of evenings and weekends. Cooldowns stop the same problem from alerting over and over.

Each organisation's servers, users, and alerts are strictly isolated from every other's, which suits an MSP looking after many clients as well as a single business. Users are admins, operators, or viewers, admins must sign in with an authenticator app, and password resets are confirmed by a WhatsApp code.