Why host Prometheus on a VPS?
Prometheus collects metrics as time series using a pull model: it regularly queries your services exposing a /metrics endpoint, which makes it particularly well suited to dynamic infrastructures and containerised environments. Hosting Prometheus on your own VPS means your metrics stay private, you can configure long retention without extra cost, and you get minimal latency between your infrastructure and your monitoring tool. Unlike SaaS solutions, a VPS gives you full customisation of alert rules, scrape intervals and integrations with your business exporters.
What you can do with Prometheus
- Monitor CPU, memory, disk and network usage of your servers via node_exporter
- Monitor your web applications and APIs by collecting custom metrics through the Prometheus clients (Go, Python, Node.js, PHP…)
- Configure precise alerts with Alertmanager to receive notifications by email, Slack or PagerDuty
- Visualise your metrics with Grafana dashboards connected to Prometheus as a data source
- Monitor your Docker containers and Kubernetes clusters via cAdvisor and kube-state-metrics
- Analyse trends over time thanks to configurable retention (15 days by default, extendable) and the PromQL query language
Install Prometheus on your ServOrbit VPS
Order a ServOrbit Cloud VPS
Head to /vps-cloud and choose the configuration suited to your load. For Prometheus alone (up to a few dozen scraped targets), a VPS with 2 vCPU and 2 GB of RAM is enough to get started. Complete the order and wait for the email confirming your VPS activation.
Open the Marketplace in your client area
Log in to your ServOrbit client area, then navigate to the 'My VPS' section and select the VPS you have just ordered. In the management panel, click the 'Marketplace' tab to access the catalogue of applications available as 1-click installs.
Launch the Prometheus installation
In the Marketplace, search for 'Prometheus' and click the application's tile. Check the information (installed version, dependencies) then click 'Install'. The system automatically provisions Prometheus on your VPS: installing the binaries, creating the systemd service, and initial configuration with default parameters.
First access to the Prometheus interface
Once the installation is complete, Prometheus is accessible in your browser at http://<YOUR-VPS-IP>:9090. The built-in web interface (Expression Browser) lets you run PromQL queries directly. By default, Prometheus scrapes itself on the endpoint http://localhost:9090/metrics — check that the target appears as 'UP' under Status > Targets.
Initial configuration and adding your targets
The main configuration file is located at /etc/prometheus/prometheus.yml. To monitor your servers, add jobs in the scrape_configs section by entering the IP addresses and ports of your exporters (node_exporter listens on port 9100 by default). After each change, reload the configuration without restarting using the command: curl -X POST http://localhost:9090/-/reload
By default, Prometheus includes neither authentication nor TLS encryption on its port 9090. In production, never expose it directly on the internet: place an Nginx or Caddy reverse proxy in front of it with HTTP basic authentication and a TLS certificate, or restrict access to port 9090 through your ServOrbit VPS firewall rules to allow only trusted IPs.
The official documentation
For advanced configuration, tool-specific options and version changes, refer to the official Prometheus documentation. This guide covers going live on a ServOrbit VPS; the vendor documentation remains the reference for fine-tuning.