Skip to content
Start monitoring free
Linux serversHow it worksFAQ
Sign inGet started

Use case

Monitor Linux CPU, RAM and disk pressure alongside website uptime. Install a small agent with systemd or Docker Compose, follow metric history, and receive alerts for sustained thresholds or missing reports.

Get started free

DEMO DATA

CPU pressure on an example host

Degraded — sustained CPU pressure

Illustrative reports · CPU threshold 90%
ReportCPU
193%
294%
396%

Three consecutive CPU reports at or above 90% activate an incident. Delivery timing depends on your notification channel.

The problem

  • A website can still return HTTP 200 while its host runs short of memory or disk space.
  • A single CPU spike should not create the same alarm as sustained pressure.
  • A stopped agent must be visible instead of leaving old values looking current.

How Pulsy helps

  • Collects host CPU, available-memory RAM usage, swap when present, filesystems, load averages and uptime.
  • Tracks CPU, RAM and each filesystem independently: three violating reports activate pressure, two healthy reports clear it.
  • Uses degraded status for resource pressure and down status after three reporting intervals without accepted reports.
  • Shares your existing dashboard, incidents, maintenance windows and notification channels. Public pages show status without private host metrics.

  1. Create a server monitor, save its one-time credential, then run the provided systemd or Compose installation command on the Linux host.
  2. Start with 90% CPU, RAM and filesystem thresholds. Choose excluded filesystems and use the effective reporting interval shown in setup.
  3. Configure a notification channel, send an explicit test, and review the first recorded result.

Pressure requires three violating reports and recovery requires two healthy reports. Missing reports become an outage after three intervals; scheduler and delivery timing also apply.

Example setup

Illustrative inputs and responses — not a live check or an import file.

Linux amd64 / arm64
CPU: 93% → 94% → 96%
RAM: 42% · /: 65%
HTTPS → /api/agent/v1/metrics

Illustrative metrics, not a command or live report. The installer uses your configured Pulsy HTTPS origin and requests the credential without echoing it.

Linux amd64 and arm64, using systemd or Docker Compose with read-only host mounts. Windows and macOS agents are outside version 1.

No. RAM alerts indicate sustained memory utilization based on available memory. They do not confirm OOM kills, inspect processes or execute commands.

Every 60 seconds by default, raised when the effective Cloud plan requires a longer minimum. A new monitor stays pending until its first accepted report. Metrics follow the effective retention policy.

Available on every Cloud plan. Each server uses one monitor slot. Reports default to 60 seconds, subject to the plan’s minimum interval; history follows plan retention.

Start with one website or Linux server. Understand the incident and know what to tell your customers.

Get started freeView on GitHub