vpsmon Cloud is a hosted monitoring dashboard for personal VPSs, homelabs, and hobby projects. It collects CPU, memory, and disk readings from Linux servers and displays them together in a single fleet view, alongside a server list showing each machine's current CPU, memory, and disk percentages and an overall server health signal. The site describes itself as "Linux VPS monitoring, without the ops team," and its headline reads "A few servers. One dashboard. $4.99 a month." It is designed for people running a small number of servers who want visibility into resource usage and clear alerts when something crosses a limit or a machine stops reporting.
The product is aimed at a specific gap: people who run a small number of servers — personal VPSs, homelab machines, hobby projects — rather than an operations team with dedicated monitoring infrastructure. The tagline "Monitor your VPSs and homelab with no inbound ports" captures the constraint the product is built around. Traditional monitoring setups typically require something to connect into the server, which means exposing a port and managing access. vpsmon Cloud inverts that: the agent on each server only sends metrics out over HTTPS, so nothing needs to be opened on the server's firewall. The result is a hosted dashboard that answers two questions the site names explicitly — "Find out, then find out why" — first whether something is wrong, then what was happening when it went wrong.
Getting a server connected is intentionally short. The documented flow is: create an account and choose Add server, run a command on the server and paste the one-use token, then watch its first readings arrive in the dashboard. The install command is a single curl piped to sudo bash, pointed at the vpsagent release on GitHub with the Cloud URL. The agent is open source — the site links to the code and invites you to "Read the code before you run it." The installer checksum-verifies the release binary first, and the agent runs as a systemd service that starts at boot. It is available for x86-64 and ARM64. Because the agent only sends metrics out over HTTPS, no inbound port is required on the server.
The dashboard's core view is a fleet table: one row per server with CPU, memory, and disk percentages, plus a count of how many servers are reporting. The site lists the tracked signals as CPU usage, memory, disk space, and server health. Beyond the at-a-glance numbers, opening an individual server shows its live processes — the top five processes by CPU and memory — so you can see what is actually running on a busy machine "without reaching for SSH first." Alerts surface inline in the dashboard as well: in the sample fleet, an incident banner reads "edge-03 · Disk limit reached," with 92% used against a 90% limit and an incident opened. The demo dashboard is also described on the site as including a missing-updates alert, showing that the fleet view surfaces more than raw percentages.
Alerting is built with defaults rather than a blank slate: the site describes "alerts with sensible defaults" and says you can change any limit. When a reading crosses its limit, or when a server stops reporting, the product notifies you by email, Slack, Discord, or your own webhook. Notifications are tied to incidents: you are told when an incident opens and when it clears, and the dashboard offers an incident timeline you can review. The site's worked example follows a disk alert in practice: agents report in, readings arrive (web-01 reports 18% CPU use), and then edge-03 reaches 92% disk use, above its 90% limit — which opens the incident and triggers the alert. That example is deliberately mundane, which is the point: the alert fires on the threshold you set, on the server you care about, without anyone watching a terminal.
Each server keeps 30 days of history. The history view is meant to answer the "why" after the "what": you can see when a spike started and what else moved with it, then switch the chart to a table of recorded readings when you want the actual numbers rather than the shape of the curve. The live demo lets you open a sample server's history directly — the demo needs no sign-up and no email — and the sample dashboard shows disk usage over the last 24 hours with a marked limit line, so the relationship between the trend and the threshold is visible at a glance.
The overall approach is deliberately minimal: a small open-source agent on each server, an outbound-only HTTPS path to the hosted service, and a dashboard that shows the fleet, the per-server detail, and the incident history in one place. There is no inbound port to open and no heavyweight stack to maintain inside the server for the monitoring layer, since the dashboard itself is hosted. The agent is installed once and starts at boot as a systemd service, and setup runs from account creation to first readings in three documented steps.
The stated benefits follow from that design. Operators of small fleets get one place to see CPU, memory, disk, and server health rather than checking each box individually; they find out about a full disk or a silent server by alert instead of discovering it later; and they can move from the symptom to the cause using history and live process data. The no-inbound-port design also means monitoring does not require opening access to the servers it watches, and the open-source agent means the code that runs on your machine can be reviewed before installation.
Concrete scenarios described on the site include a disk filling on an edge server — the demo's edge-03 sits at 92% against a 90% limit with one active alert — and a busy web server where you want to see the top processes by CPU and memory before opening an SSH session. A homelab or a set of hobby VPSs is the stated target: the sample fleet includes web-01, web-02, db-01, worker-01, and edge-03, and the base plan is scoped to up to five servers. Reviewing an incident timeline after the fact, checking 30 days of recorded readings in table form, and catching a server that has stopped reporting are all directly supported. For a single server, the separate open-source vpsmon project offers a self-hosted dashboard instead.
Pricing is stated plainly. The Cloud base plan is $4.99 USD per month for up to five servers, with a 7-day free trial and no card required to start. The price includes applicable transaction taxes and renews monthly until canceled, with billing and refund terms published on the site. The product is explicitly scoped for personal, non-business use only. For a single server, the open-source vpsmon project is a separate, self-hosted dashboard with a $0 software license: it runs on your own server, keeps metrics on that server, and requires no Cloud account. Installation docs are published, along with a guide on how to monitor a Linux VPS.
For anyone running a few Linux servers without an ops team, vpsmon Cloud's value proposition is narrow and clear: one hosted dashboard for CPU, memory, disk, and server health across up to five machines, alerting by email, Slack, Discord, or webhook when a limit is crossed or a server goes quiet, 30 days of history to explain what happened, and an open-source, outbound-only agent that never asks you to open an inbound port.