Case studies: what we built, what we run, what we measured

This page lists work whose results we can back up: each with its scope, the approach and what it does not include. The first case study is our own product, Fortlauf. We built it, and we run it on our own hardware in a data centre in Munich.

Our own operations

Every measurement from our own operations, with its date and scope, is on Infrastructure: where your data lives.

How we write case studies

  • Anonymous unless the client agrees otherwise. We describe client work without names and without details that would identify the client.
  • Numbers only where measured. We give a number only where it was measured, and say what it was measured on and what it leaves out.
  • Checked before published. A client case study appears here once its sources and its release have been checked.

What you will not find here

Client logos used without permission, quotes without consent, and numbers that were not measured.

Case studies with measured results

Our own product

Fortlauf: running a SaaS on our own infrastructure

Fortlauf is our product. We built it, and we run it on our own hardware. In production since July 2026: tenant isolation in the database, immutable document storage, delivery by GitOps, measured recovery.

Company with its own Ceph cluster

Client Ceph cluster encrypted, zero downtime

A customer wanted its Ceph cluster encrypted after the migration. We encrypted every OSD while the cluster stayed in service, with zero downtime, and handed over secrets, documentation and runbooks.

Small manufacturer of safety systems

Hardware maker: one process for code and testing

A manufacturer of safety systems introduced a Kanban development process covering code, designs and hardware tests. We advised on it and set up GitLab, GitLab CI, OpenProject and paperless-ngx on their own VMs with SMTP and LDAP/SAML2.

Connected-device vendor

IoT audit: MQTT vulnerability found and fixed

Full-stack security audit for a connected-device vendor: a critical MQTT vulnerability would have allowed device takeover. We helped fix it through a secure firmware update process, plus continuous scanning of containers and package registries.

Logistics SaaS for plant and transport logistics

Logistics SaaS: p95 latency cut by about 80 %

While onboarding about 50,000 users, a logistics SaaS platform ran into timeouts. We analysed the edge logs automatically, re-sized resources and autoscaling on EKS and added the missing indexes. The p95 response time fell by about 80 %, compared over one month before and one month after the changes.

Company with an on-site data centre

Company-wide LLM platform in a client data centre

Three GPU servers with NVIDIA A100 sourced and brought online in the customer's data centre: Kubernetes with GPU nodes, vLLM for inference, LiteLLM as the company-wide proxy. Then firewalls and switches made redundant and proven with failover tests.

Passenger information at a large transport operator

Passenger information: release tests automated

End-to-end tests of an event-driven data platform (Kafka, Spring Boot): we rebuilt the test orchestration from plain GitLab CI jobs to Argo Workflows, provisioned release environments via Crossplane on Kubernetes and refactored the test runner.

Working on something similar?

Planning a platform, a SaaS product or a move onto your own hardware? Tell us what it is about. All six services are listed under Services.