Each server you add gets its own page with live charts, health details and history. The agent reports every 30 seconds by default. Change it with Report every in the server's Settings (10, 15, 30 or 60 seconds).
What the agent collects
| Metric | Details |
|---|---|
| CPU usage | Percentage across all cores, and the usage of each core. |
| Load average | 1, 5 and 15-minute load averages (alerts use the 1-minute value). Compare them with the number of cores. |
| Memory and swap | Percentage in use, plus totals. |
| Disk and inodes | Usage for every mounted filesystem. Disk alerts and the forecast follow the root filesystem (/); inode alerts follow the fullest one. |
| Disk activity | Read and write speed, and how busy the disk is. |
| Network | Bytes in and out per second. |
| Services | Whether the systemd services and processes you choose are running. |
| Containers | State, health, CPU and memory of Docker containers (with --docker). |
| Processes and ports | The 10 busiest processes by CPU and the 10 using the most memory, plus up to 100 listening ports. |
| Package updates | How many updates are waiting, how many are security updates, and whether a restart is needed. Checked every 6 hours. |
Server status
| Status | Meaning |
|---|---|
| Waiting for agent | The server was added but the agent hasn't reported yet. |
| Online | Reporting normally and no alerts are open. |
| Warning / Critical | An alert with that severity is open, from a rule or from a check like Service down or New open port. |
| Offline | No report for three reporting intervals (at least one minute). You get an alert. |
| Paused | You paused monitoring. No alerts are sent. |
Charts and history
Charts show the last hour, 6 hours, 24 hours, 7 days or 30 days. How far back you can go depends on your plan's history setting. Older samples are deleted automatically.
The server page also shows:
- A health score out of 100 (Healthy, Needs attention or Unhealthy), with what cost it points.
- System details (OS, kernel, CPU cores, uptime), every disk and network interface, and bandwidth used today, this month and over the last 30 days.
- Notices such as pending updates, a restart being needed, a firewall that's off, zombie processes, out-of-memory kills and recent sudo users.
- Availability over the last 30 days, and the 8 most recent alerts with how long each one lasted.
- Alert rules that apply to this server, with a shortcut to add one just for it.
- Upcoming maintenance that covers it.
- Team notes, for things like "runs the billing cron at 02:00".
- An AI health check, if your team uses AI features.
Tags and settings
Give servers tags like production or db (up to 10 each) to filter the server list and aim alert rules at a group. The server list can also be searched by name, hostname or IP address.
Open Settings on a server to rename it, change its tags, set Report every, generate a new agent token or delete it.
Disk forecast
After about 12 hours of data, Alertly.Net looks at the last week of disk usage on / every hour and estimates when it will be full. If that's within a week, you get a Disk filling up alert with the expected date (critical when it's within a day), long before it becomes an outage. The server page shows the estimate even when it's further off.
Pausing a server
Use Pause on the server page when a machine is switched off on purpose. While paused, new metrics are ignored and no alerts are sent until you resume. For scheduled work, a maintenance window is usually better because it ends on its own.
Removing a server
Uninstall the agent first (see Installing the agent), then open the server's Settings and choose Delete server. Its metrics, alerts, notes, the rules made just for it and its status page components are deleted with it.