Server Monitoring

Install one tiny Node script (or single Docker container) on any Linux server. Reports CPU, memory, disk, network, and load average to FlowAPI every 30 seconds with threshold alerts and a 7-day time series.

Get Started Free

What You Get

Stdlib-only agent

Single-file Node script under 300 lines. No npm install, no agent registry. Read-only access to /proc and /.

What it measures

CPU%, memory used/total, swap, load average 1/5/15 min, network rx/tx per-second, disk per-mount usage with /-rollup.

Thresholds + alerts

Per-server thresholds for CPU%, MEM%, and disk%. Alerts on edge transitions only — no spam.

Sparklines + cards

Live stat cards with last-hour sparklines. Red border when over threshold.

Disk breakdown

Per-mount usage table with progress bars; flags individual mounts that exceed the disk threshold.

No Docker required

Install via systemd on bare Linux. Docker is optional — works on Pi, ARM VPS, anywhere Node 18+ runs.

Frequently Asked Questions

What does the agent need on the box?

Node 18.15+ and read access to /proc and the root filesystem. The install command installs Node 20 from NodeSource if your distro's Node is older.

Does it read application data, logs, or env vars?

No. It only reads /proc/* (CPU, memory, network, load) and walks filesystem mounts for usage. The source is under 300 lines — auditable in one sitting.

How is this different from Datadog Infrastructure?

No per-host pricing. Single tiny agent. Threshold alerts route to the same notification channels (Slack / Teams / email / webhook) as your API monitors.

Retention?

Defaults to 7 days of metric samples. Configurable via the SERVER_METRICS_RETAIN_DAYS env var.