Grafana Hosting on Your Own VPS
Dashboards and alerting over your own metrics, without per-seat or per-metric billing.
Size a server for Grafana
Every stop is a real configuration you can deploy. Priced per 30 days, billed by the hour, exclusive of 18% GST.
What Grafana is
Grafana is a dashboard and alerting layer. It does not store metrics itself - it queries data sources such as Prometheus, InfluxDB, Loki, Elasticsearch or a plain SQL database, and turns the results into panels, dashboards and alert rules.
That distinction matters when sizing a server. Grafana itself is light: it renders queries and holds dashboard definitions in a small database. What is heavy is whatever it queries. A VPS running only Grafana can be modest; a VPS running Grafana alongside Prometheus needs to be sized for Prometheus, whose memory use scales with the number of active time series.
Self-hosting removes two limits that hosted observability imposes. Seats - hosted plans typically bill per user, which discourages giving a dashboard to everyone who would benefit. And ingestion - hosted plans meter data volume, which quietly encourages collecting less than you should.
Why run Grafana on a VPS
No per-user pricing, so read access can go to the whole team.
No ingestion metering - retention is a disk decision, not a billing one.
Grafana and its data sources on the same private network, so queries do not cross the internet.
Root access for plugins, rendering and alert delivery configuration.
Grafana features
Dashboards over many data sources at once
Alerting with routing to email, chat or webhooks
Prometheus, Loki, InfluxDB, Elasticsearch and SQL data sources
Templated dashboards with variables
Teams, folders and permissions
Rendered PNG panels for reports
API and provisioning for dashboards as code
Recommended server
A starting point for Grafana, not a hard floor — size it on the sliders.
| Memory | 2 GB for Grafana alone; 8 GB or more if Prometheus runs beside it |
|---|---|
| CPU | 2 vCPU; more if many dashboards refresh at once |
| Storage | 40 GB for Grafana; metric retention is what really consumes disk |
| Operating system | Ubuntu 22.04 LTS or Debian 12 |
How deployment works
Grafana is installed as a service behind a reverse proxy with TLS for your domain. Its own database can stay SQLite for a single instance or move to PostgreSQL. If you want Prometheus or Loki on the same machine, size for those rather than for Grafana - they are what consume the memory.
What people run it for
Infrastructure monitoring
Host and container metrics with alerts on saturation, not just on outage.
Application dashboards
Latency, error rate and throughput per service, correlated on one screen.
Business metrics
Query a production database directly for orders or signups alongside system health.
Log correlation
Loki beside metrics, so a spike and its logs are in the same view.
Grafana hosting questions
Related applications
Zabbix Hosting Agent-based infrastructure monitoring with no per-host licence and retention you decide. 
Uptime Kuma Hosting Uptime monitoring and status pages, with unlimited monitors and no per-check billing. 
PostHog Hosting Product analytics and session recording with event data that never leaves your server. 
Elasticsearch Hosting Full-text search and log analytics, with heap and retention under your control. 
Windmill Hosting Scripts become APIs, workflows and internal UIs — with executions bounded by your server, not a plan. Deploy Grafana on your own server
Root access, your choice of size, and no container between you and the application.