Skip to content

Managed GitLab Runner

Runners for your GitLab pipelines on dedicated hardware in Frankfurt am Main. Docker with dind is set up, the cache is included, and we install the updates.

The runners in brief

What teams want to know before they switch their pipelines.

Autoscaling?Yes, by capacity.More concurrent jobs per runner and additional runner instances as needed. Classic cloud autoscaling is deliberately not the default.
Location?Frankfurt am Main.Dedicated hardware, ISO 27001 and GDPR-compliant, DPA on request.
Your own runners?Yes, in parallel.Alongside our managed runners you can register your own runners at any time, for example for special hardware or local tests.
Caching?Up to 100 GB included.Free cache per runner. Dependencies and intermediate results are reused, which shortens build time.
Executor?Docker with dind.Docker-in-Docker is the default, plus preconfigured packages for container builds. Kubernetes executor on request.
Network?From 1 Gbit/s.If your GitLab instance also runs with us, runner and server are connected at 10 Gbit/s.

More about the data center in Frankfurt am Main where your instance runs.

Runner overview of a GitLab instance with three runners maintained by NETZFABRIK in Frankfurt, all online

Which executor do you run?

The default is Docker with dind. We set up Kubernetes on request, and dedicated macOS runners are available too.

Default

Docker & dind

Full Docker-in-Docker support. You build, tag and push images directly in the pipeline.

On request

Kubernetes

We set it up on request. The default stays Docker with dind, because most teams build and test images that way.

from 99 €

macOS

Dedicated macOS runners with a preinstalled Xcode environment for iOS and macOS builds.

What happens between push and result

Nothing changes in your pipeline. The only difference is that we look after the runners.

pipeline · zsh

ops@netzfabrik~/pipeline% git push origin main

  1. → 01Commit triggers itA push or merge request starts your .gitlab-ci.yml pipeline in GitLab, exactly as usual.
  2. → 02GitLab dispatches jobsGitLab hands the jobs to the matching runners via tags. Your managed runners report in with free capacity.
  3. → 03Runner executesEvery job runs isolated in the Docker executor with dind: build, test, lint and container push.
  4. → 04Results & cache returnArtifacts and status go back to GitLab, the cache stays warm. We monitor load and errors in the background.

Official GitLab Open Partner

We are a registered GitLab Open Partner. If your instance needs Premium or Ultimate, you get the licenses directly from us.

Partner certification: GitLab Open Partner

Is it worth the effort?

Runners are quick to register. But patches, executor upkeep, cache and autoscaling cost ongoing time. We take over operations at clear package prices.

Is it worth the effort?
Self-managedNETZFABRIK Managed
Setup & registrationSet up runners and executor yourselfSet up and included
Executor & Docker-in-DockerSet up and maintain dind yourselfDocker + dind ready, we maintain it
CachingSet up and manage cache storage yourselfUp to 100 GB per runner included
Updates & patchesKeep runner updates on your radarAutomatic, by us
MonitoringMonitor runner health yourself24/7 included
Scaling & autoscalingPlan capacity and autoscaling yourselfBy capacity, we add runners
LocationLocation is your responsibilityFrankfurt am Main, ISO 27001, GDPR
Your time investmentSeveral hours per monthAlmost none, we operate it

Runner plans by compute power

Monitoring and caching come with every plan. The larger ones add more cores, more RAM and extra S3 cache.

1. Choose a plan

2. Have GitLab hosted as well

Your builds need more cores or more cache? Then we put together a runner on dedicated hardware based on your jobs.

Have GitLab hosted as well

Need a home for your GitLab instance as well as runners? We run GitLab as a managed service, including updates, monitoring, backups and support.

If GitLab and runners both run with us, they are connected at up to 10 Gbps. Artifacts and images then do not travel over the public internet.

Response times for your GitLab Runners

In the event of a system outage, we respond in under 15 minutes, 24/7, 365 days a year.

  • Critical (P1)< 15 min

    Total outage or suspected security incident. Please call.

    24/7 on-call

  • High (P2)< 1 hr

    Important function disrupted, operations run with limits.

    Mon to Fri, 9 am to 6 pm

  • Normal (P3)< 4 hrs

    One workstation affected or there is a workaround.

    Mon to Fri, 9 am to 6 pm

  • Low (P4)< 1 business day

    Questions, requests and anything that can be planned.

    Mon to Fri, 9 am to 6 pm

Questions about the managed runners

First the technology, then performance, scaling and location.

Technology & operations

Which executor types are supported on your Managed GitLab Runners?
By default we use the Docker executor with full Docker-in-Docker (dind) support. This lets you build, tag and push Docker images directly in your CI/CD pipelines. In addition we provide preconfigured packages that further simplify the build process for container images.
Do you also offer macOS runners?
Yes. For teams building iOS or macOS apps we offer dedicated macOS runners from 99 euro per month. They come with a preinstalled Xcode build environment, and we handle maintenance and updates. Swift and Objective-C builds then run directly in your GitLab CI/CD pipeline.
Do you also offer the Kubernetes executor?
The default is the Docker executor with Docker-in-Docker, because most teams build and test images that way. We set up the Kubernetes executor on request if you want jobs to run in an existing or NETZFABRIK-operated cluster. For that we need your target cluster setup up front.
Does my GitLab have to be hosted with you to use your runners?
No. You connect our managed runners to any GitLab instance: hosted with us, at another provider or on gitlab.com. If your GitLab also runs with us, runner and server are connected internally at 10 Gbps, so large artifacts transfer faster.
How does caching work with the GitLab Runners?
Each managed runner includes up to 100 GB of free caching storage. Build artifacts, dependencies and intermediate results stay there between pipeline runs, so not every run has to fetch everything again. If you need more space, we expand the quota on request.

Performance, scaling & location

How fast is the network connectivity of the runners?
Each runner is connected with at least 1 Gbps. If your GitLab instance also runs with us, a dedicated 10 Gbps connection is available between runner and server. You notice it most with large repositories, big artifacts and container registry pushes.
Can I adjust runner resources as demand grows?
Yes. Per runner we adjust CPU, RAM, and local storage, for example for heavier Docker builds or more concurrent jobs. We can add runner instances quickly when your pipeline queue grows and scale them back down after the peak.
Do you support autoscaling for GitLab runners?
For peak load we scale capacity: more concurrent jobs per runner and additional runner instances as needed. Classic cloud autoscaling via Docker Machine or Fleeting is not a standard package. If you need dynamic runner pools, we design a suitable setup with you.
Where do the GitLab runners run?
Runners run on dedicated hardware in Frankfurt am Main, ISO 27001 and GDPR-compliant. If your GitLab also runs with us, job traffic uses the internal high-speed path instead of detours over the public internet. We provide a DPA on request.
Can I additionally connect my own runners?
Yes. Alongside our managed runners you can register and run your own runners in parallel at any time, for example for special hardware requirements or local test environments. Configuration is done directly via the GitLab Runner administration of your instance.

Move your pipelines to our runners?

Tell us which pipelines run and where your GitLab lives. We set up runners with Docker, dind and caching to match.