Service 01 / 04
Design & setup
An off-the-shelf cluster rarely fits. We design your environment starting from your application: load profile, availability target, and budget determine the architecture – not the other way around.
Approach
From load profile to architecture.
It does not start with a hardware list – it starts with a conversation: What runs on the cluster, how many users access it, how much downtime is tolerable, and how much is not? From the load profile, availability target, and growth expectations we derive the architecture: number and sizing of nodes, storage concept, network layout, and degree of redundancy.
We plan deliberately soberly: as much redundancy as necessary, as little complexity as possible. Every additional component has to justify its place in the concept – what does not exist cannot fail and does not need maintenance.
Technologically, we rely on proven, open components – for example virtualization with Proxmox VE and KVM, replicated storage, load balancing with HAProxy and keepalived, and database replication for PostgreSQL or MySQL/MariaDB. What exactly gets used is decided by your use case, and it is spelled out clearly in the concept.
Scope
What the design phase includes.
- Assessment & requirements analysis: application, load profile, availability and recovery targets (RTO/RPO).
- Sizing of CPU, RAM, storage, and network, including headroom for load peaks and growth.
- Redundancy concept: N+1 design, quorum planning, and protection against split-brain situations.
- Virtualization environment: setup and configuration, e.g. with Proxmox VE / KVM.
- Load-balancing concept with redundant load balancers and proper health checking.
- Network architecture including VLAN separation, firewalls, and redundant connectivity.
- Storage & replication concept matched to data volume, IO profile, and RPO target.
- Migration plan for moving your application, with a minimal and predictable cutover window.
- Failover test before handover: we pull the plug before chance does.
- Documentation: architecture, configuration, and runbook – in writing and understandable.
Reference
A typical target architecture.
Process
From the first conversation to a running cluster.
-
1
Assessment
We review your application, load data, and existing infrastructure, and clarify availability targets and budget.
-
2
Architecture & fixed price
You receive a written concept with an architecture drawing, component list, and a fixed price for setup and operations.
-
3
Setup & failover test
We build the cluster, configure monitoring and backups, and test every failover scenario before any load runs on it.
-
4
Migration & handover
Your application moves in a controlled way. Then day-to-day operations take over, with documentation and a dedicated contact.
FAQ
Common questions about the design phase.
Do we need our own hardware or data center?
No. The cluster runs on infrastructure we manage in German data centers. If you already own hardware or colocation space, we are happy to check whether and how it can be integrated.
How long does it take from inquiry to a finished cluster?
Concept and quote are usually ready within a few business days. Setup and testing typically take one to three weeks depending on scope; you choose the migration date.
What does a managed cluster cost?
That depends on sizing and redundancy level, which is why we quote prices only after the concept – but then as a fixed monthly price with no hidden hourly rates. The initial assessment is free of charge.
Can you take over an existing environment instead of building new?
Yes. We audit the existing environment, document deviations from the target design, and take over operations – either as-is or with a step-by-step rebuild.
Contact
Tell us about your application.
You will receive an initial assessment including an architecture idea within one business day, free of charge and without obligation.
Request a project