Hetzner added the CPX02 on October 6, a Regular Performance cloud server with 1 vCPU, 1 GB of RAM and 20 GB of NVMe for static sites, blogs and small single apps. On October 12 it was listed as currently unavailable, like the rest of the CPX line.
The New Plan
Hetzner added the CPX02 to its Regular Performance line on October 6, 2026.
- 1 vCPU on AMD, 1 GB of RAM and 20 GB of NVMe storage
- Falkenstein, Nuremberg and Helsinki in Europe, plus Singapore
- 20 TB of traffic a month in the European locations and 0.5 TB in Singapore
Hetzner describes it as a plan for workloads with very low resource needs, such as static websites, blogs and simple web applications that run a single service. Regular Performance servers share their CPU, and Hetzner says they run on newer hardware generations than its Cost-Optimized plans, which gives them higher overall CPU performance.
What Fits in 1 GB
1 GB is plenty for software that does one thing well.
- A static site or documentation behind Nginx or Caddy
- A WireGuard VPN endpoint or a small reverse proxy
- An uptime checker or a status page
- A small blog. WordPress can run too, but PHP and MySQL share the same 1 GB, so add a swap file and page caching before traffic arrives.
You May Not Be Able to Order It Yet
On October 12, Hetzner's Regular Performance page showed the CPX02, and every other CPX plan, as currently unavailable, with a note to check back later. Hetzner's status page has carried a notice since June 26, 2026 explaining that cloud servers can only be provisioned to a limited extent because of high demand and limited hardware. New servers are restricted for new customers and for some existing customers chosen at random, and individual requests cannot be prioritised.
Servers that already run are not affected. Two habits help while the notice stands.
- Check the Cloud Console for your project and location before planning a launch around the CPX02.
- Think twice before deleting a server you may want back, since creating a replacement might not be possible right away.
Also New This Week
- Load balancers can now use an existing Primary IP instead of having an address allocated automatically. The IP is chosen when the load balancer is created and cannot be changed afterwards.
- Hetzner has scheduled maintenance on the emergency power systems in Helsinki, for HEL1-DC1 on October 13 and HEL1-DC2 from October 14 to 15, and maintenance work on load balancers on October 13.
Where It Fits
Hetzner's plans and prices have moved a lot this year, as our note on the June cloud price changes shows. Sizing servers and keeping a fleet running through changes like these is everyday work for our cloud management team.
Sources
Or read how we handle it in Managed Cloud.
Related News
Hetzner More Than Doubled Some Cloud Prices Today And What To Do About It
On June 15, 2026, Hetzner repriced its cloud servers, and the dedicated and AMD shared vCPU lines (CCX, CPX) jumped 113 to 175 percent while the ARM and Intel-shared lines rose about 30 percent. Existing instances are protected, but a rescale reprices you. Here is exactly what changed and what to do.
CloudPlanning a DigitalOcean Database? Read This Before Oct 15
From October 15, new DigitalOcean accounts can no longer create Standard Edition PostgreSQL or MySQL clusters above 4 GiB of RAM or with standby nodes, and from November 30 the same applies to every account. Existing clusters keep everything.
CloudAWS Shuts Down App Mesh for Good on September 30 and Your Service Mesh Goes With It
AWS ends all support for App Mesh on September 30, 2026, after which the console and every App Mesh resource go dark and the Envoy data plane is no longer managed. New onboarding has been blocked since September 24, 2024. For ECS, AWS points to Service Connect; EKS teams have their own path.