One lightweight agent for Linux, macOS and Windows. CPU, memory, disk and network every 30 seconds, for a fraction of Datadog’s price.
1 server and 20 monitors free. No credit card required.
User, system and iowait time as a share of wall time, with 1, 5 and 15-minute load averages. On Windows, load comes from the processor queue.
Used, available and total memory, and utilization as a percentage. A leak shows up as a slope days before it takes the host down.
Used and free space for every real mount point, with pseudo-filesystems left out. On Windows, each drive letter is a mount.
Bytes read and written per block device, to find the drive that saturates when requests start to queue.
Bytes in and out per interface, turned into download and upload rates in the dashboard. Loopback and virtual adapters are left out.
Hostname, OS, kernel, architecture, CPU model, boot time and agent version, current on every scrape. The dashboard flags agents with an update waiting.
Servers sit in the same project as your monitors, alert through the same escalation policies and page the same on-call.
An uptime monitor tells you the API stopped answering. The agent on the host behind it shows the memory it ran out of, in the same project.
Uptime monitoringCron job monitoring catches the job that didn’t check in. The disk space chart of the database host shows why it couldn’t.
Cron job monitoringWhen a server stops reporting, an outage opens and its escalation policy pages whoever is on call, exactly as it does for your monitors.
On-call schedulingMetrics arrive on every scrape. Liveness comes from them, there is no separate heartbeat.
Flagged in the dashboard only. A network blip doesn’t wake anyone up.
An outage opens and the server’s escalation policy pages its channels. The threshold is yours to tune, from 60 seconds.
The outage resolves and the same channels get the recovery notification.
Bind an escalation policy to each server. Its steps, delays and channels apply as they do for your monitors.
Slack
Microsoft Teams
Google Chat
PagerDuty
Opsgenie
Jira Service Management
Discord
Telegram
Mobile appIn the dashboard, open Servers and click New server. Hyperping gives you the install command with your project’s token.
With sudo on Linux and macOS, in an Administrator PowerShell on Windows. The installer checks the download’s SHA256 and registers a service.
About 30 seconds later the server turns Online and its CPU, memory, disk and network charts start to move.
# Linux (systemd) or macOS (launchd), amd64 or arm64, with sudo
curl -fsSL https://hyperping.com/install.sh | sh -s HP_INSTALL_xxxxxThe agent targets about 50 MB of memory and scrapes every 30 seconds, next to your app rather than on a monitoring box.
Unsent batches wait in a queue on disk and are retried, so graphs stay whole through reboots, network blips and ingest outages.
An embedded OpenTelemetry collector reports standard hostmetrics names over OTLP/HTTP. No collector of your own to run.
The agent only calls out over HTTPS. Nothing listens on the host and nothing needs opening in your firewall.
The install token is swapped for a token of the host’s own. Re-run the installer to rotate it, or revoke it from the dashboard.
Set the name, escalation policy and group at install time from Ansible, Terraform or a DaemonSet. Re-installs keep the edits made in the dashboard.
Infrastructure monitoring, billed per host.
$300/month
$3,600 over 12 months
Servers, uptime, status pages and on-call.
$74/month
$890 billed yearlySave $2,710 a year on 20 servers, before Datadog’s uptime checks and on-call.
Reviewed . All amounts are USD per month on annual billing; local currency prices may differ.
Linux with systemd and macOS with launchd, on amd64 and arm64, and Windows 10 or Windows Server 2016 and later on amd64. Linux and macOS use a one-line shell installer, Windows a one-line PowerShell installer that registers a Windows service. See Install the agent for the details.
The Free plan includes 1 server agent with 2 days of metric history. Essentials includes 5 agents, Pro 20 and Business 100, with server alerting on every paid plan. Extra agents are $2 a month, or $1.50 on Business. See pricing, or compare Datadog, New Relic and Netdata host by host in the server monitoring cost calculator.
The agent targets about 50 MB of memory and scrapes host metrics every 30 seconds. It is built to run on production machines next to your workload, not on a dedicated monitoring host.
A server turns Stale after 30 seconds without metrics, which only shows in the dashboard. Once its offline threshold passes, 90 seconds by default and 60 at the least, an outage opens and the escalation policy bound to the server pages its channels. When metrics come back, the outage resolves and a recovery notification goes out. See Server alerting.
No. The agent only makes outbound HTTPS requests to api.hyperping.io and ingest.hyperping.io. If they fail, metrics are queued on disk and sent once the connection is back.
At 1-minute resolution, 2 days on Free, 7 on Essentials, 14 on Pro and 30 on Business. Coarser rollups are kept longer: up to 6 months of hourly points on Essentials and 13 months on Pro and Business. See Metrics collected.
For host metrics, yes: install the agent on the VM or node, for example with a DaemonSet, and you see that machine’s CPU, memory, disk and network, including the load of the containers it runs. Per-container and per-process metrics are not collected.
Datadog is a full observability platform, with APM, logs and traces billed per host, per gigabyte or per event. Hyperping covers host metrics, uptime monitoring, status pages and on-call in one plan with a fixed price. If you need trace search or log retention, you need an observability tool. Read the Datadog alternative comparison.
