Skip to main content

FAQ

Inspect 1.36

What is Syself Autopilot?#

Syself Autopilot gives you highly available Kubernetes clusters on Hetzner that are ready for production, without assembling, patching, or operating the base infrastructure yourself. It is a fully managed platform: you run your workloads, Syself runs the machinery underneath.

Syself Autopilot builds on Cluster API, the open CNCF standard for cluster lifecycle, so your clusters are standard Kubernetes objects. You drive them with kubectl, and your manifests, Helm charts, and tooling work unchanged.

How do I get access and start?#

There is no self-service signup yet. For now, you get started by reaching out, which also lets us understand your situation and set you up the right way. You have two options:

  • Schedule a meeting to talk through your current setup and how we can support you. Book a call.
  • Email us at contact@syself.com⁠ with your first and last name, company, and email, and we will set you up with a 14-day trial or a free-plan account.

Once you reach out, Syself provisions your account and emails you back an invite to set a password and log in, along with the details you need to start, including your namespace. From there:

  1. Install kubectl and kubelogin. See .
  2. Prepare your Hetzner account: create a project, generate an API token, and upload an SSH key. See .
  3. Access the management cluster. See .
  4. Create your first workload cluster. See .

Can Syself help with more than running clusters?#

Yes. We also help with everything next to the cluster: best practices, integrations, observability, running databases, and migrations, whether you are moving from another provider or adopting Kubernetes for the first time. We do this directly or together with our partner network, and we can back it with an SLA. If you want help scoping any of this, email us⁠ or book a meeting.

Which Hetzner locations are supported?#

Syself Autopilot runs on Hetzner Cloud. Available regions include fsn1 (Falkenstein), nbg1 (Nuremberg), hel1 (Helsinki), ash (Ashburn), and hil (Hillsboro). You set the location once per cluster with the region topology variable (default: fsn1). Each cluster runs in one region.

What is the difference between the management cluster and workload clusters?#

Syself Autopilot uses two separate Kubernetes clusters.

The management cluster is operated by Syself as part of the service. It runs the controllers that provision, upgrade, and self-heal your clusters. You do not own or run this cluster.

A workload cluster runs your applications. Each one is a standard, upstream Kubernetes cluster. It is yours: your nodes, your namespaces, your workloads.

The two clusters are always separate. Your workload clusters keep running even when the management cluster is unreachable. Management operations (provisioning new nodes, running upgrades, replacing failed machines) stop while the management cluster is unavailable, but your running workloads continue.

See for the full breakdown.

Where does the management cluster run?#

Both options give you the same cluster management. The difference is where the management cluster runs.

By default, Syself hosts it. Your organisation gets its own namespace on a management cluster that Syself hosts and operates, so you do not run the management cluster or the identity system yourself.

You can also run it in your own Hetzner account. Syself still operates it for you, the same way it operates your workload clusters; the difference is where it runs, not who keeps it healthy. This suits organisations with stricter privacy, security, or sovereignty requirements. To inquire, email our sales team at contact@syself.com⁠.

See for more detail.

How is billing calculated?#

You pay Hetzner directly for the infrastructure. On top of that, Syself charges a management fee based on the resources you use for your managed clusters. To measure that usage, Syself reads the structured CSV file Hetzner includes in its monthly invoice. The CSV covers both Hetzner Cloud and Hetzner Robot (bare metal), and Syself filters it down to the resources in your managed clusters.

The standard setup is to add Syself's email address as a billing contact in your Hetzner account. Hetzner then sends the monthly invoice, including the CSV, to Syself automatically.

On top of the usage-based part, there is a fixed base fee that does not depend on how many clusters you run. It pays for the ongoing work behind a managed cluster, not just running one: security patching, tested version upgrades, the integration work to keep everything fitting together, and support. The version hub shows that work in the open, including the CVE tracking behind each release. The base fee matters most for smaller clusters: Hetzner is cheap and clusters can be small, so the usage-based part alone would not cover keeping them patched and secure.

A few options make it easier to start:

  • Free under 5 nodes. A cluster smaller than 5 nodes runs for free, with no support included.
  • 14-day trial. Syself Autopilot has a 14-day trial.
  • Monthly contract. The contract renews monthly. You can cancel on a monthly basis once no clusters are managed anymore.

See for full details.

How do upgrades work? Is there downtime?#

Upgrades replace nodes; they do not patch them. When you point a cluster at a newer stack version, the controllers bring up new nodes from the new sealed OS image and remove the old ones.

Nodes roll one at a time, not in parallel. The control plane rolls first, one node at a time, and etcd (the database that stores all cluster state) keeps a working majority throughout, so the Kubernetes API stays available. Each worker pool then rolls, one node at a time.

See and for the step-by-step process, including how to keep an application available during the drain.

What happens if Syself has an outage?#

Your running clusters keep serving traffic. The management cluster is separate from your workload clusters, so an outage on Syself's side does not touch your nodes, your API server, or your workloads. They run on standard, upstream Kubernetes and do not depend on Syself to stay up.

What pauses is management. While the management cluster is unavailable, you cannot create a new cluster, or scale a cluster up or down. Those operations run on the management cluster, so they wait until it is back, then resume on their own. Nothing you are already running is affected in the meantime.

How do I get support?#

Email Syself support at support@syself.com⁠. If you are working through the getting-started guide and hit problems, the covers common issues with provisioning and servers.

For user management and access questions, see and .

Related#

Grant Syself bare-metal access

Put a Hetzner bare-metal server into maintenanceMode, create a temporary Hetzner Robot admin login for Syself support, then revoke it and let the server rejoin the cluster.

How to Add Users to Your Organization

Add users to your organization from the Syself ID console.