hyggecloud / projects

What a project with us looks like, concretely

Two typical project journeys in detail — from starting point through stumbling blocks to the numbers afterwards. Anonymised example scenarios based on real constellations: industry, scale and figures show what you can expect.

Scenario 01 · Startup

SaaS platform: from AWS to Hetzner in 9 weeks

B2B SaaS (HR tech) · ~45 employees · ~30,000 active users · Series A

Starting point

Classic AWS sprawl after four years of growth: EC2, RDS (PostgreSQL), S3, CloudFront, SQS, a half-forgotten Lambda tangle. Monthly bill: ~€19,000 and rising. Trigger: two enterprise prospects demanded EU hosting without US jurisdiction in their security reviews — and the CFO had been demanding answers on cloud costs for months anyway.

The path

  • Weeks 1–2 (check): inventory: 47 services, 9 of them unused ("zombie infrastructure": ~€1,400/month for nothing). Lambda analysis: 12 functions, 10 of them trivial cron jobs.
  • Weeks 3–6 (build): k3s cluster on Hetzner (Falkenstein), Terraform + Argo CD, RDS replaced by a Patroni HA setup, S3 data (2.3 TB) synced to Hetzner object storage via rclone.
  • Weeks 7–8 (parallel operation): staging under real load, data replication at minute intervals, load tests with production traffic replay.
  • Week 9 (cutover): DNS switch on a Tuesday evening, 4-minute maintenance window for the final DB sync. Customer complaints: zero.

The stumbling block (honestly)

Two Lambda functions were deeply wired into SQS + EventBridge. Instead of forcing a replacement: rebuilt as a lean worker container with NATS — planned two weeks earlier, that would have shortened parallel operation. Lesson learned; it's in every check report now.

−67 %infrastructure cost/month
€19.0 → 6.2kmonthly bill
4 minmaintenance window at cutover
~5 mo.break-even incl. project cost

Afterwards

Handover with runbooks to the internal dev team; HyggeCloud only handles quarterly patching now. The two enterprise deals were signed, by the way — the EU hosting chapter in the security questionnaire is now one page instead of an escalation meeting.

"The moment the CFO saw the new bill was the best product demo we've ever had."

Scenario 02 · SME

Machine builder: from Azure to IONOS — with operations included

Manufacturing · ~180 employees · ERP integration, customer portal, file services · no in-house ops team

Starting point

A historically grown Azure setup, once installed by an IT reseller: VMs for ERP middleware, a customer portal, Azure Files for the CAD archive (11 TB). The trigger was the data protection officer: after the Microsoft/ICC incident he asked the question nobody could answer — "What's our plan if that happens to us?" Plus: Azure costs of ~€8,500/month that nobody could fully explain anymore.

The path

  • Check (2 weeks): a decision paper for management — including the honest finding that the ERP itself (SaaS from a German vendor) wasn't affected at all. Only what made sense was migrated.
  • Provider choice: IONOS over Hetzner — decisive were the BSI C5 attestation for their automotive supply-chain customers, a German contract partner and German phone support for the internal IT lead.
  • Migration (7 weeks): VMs rebuilt as IaC instead of lift & shift — old baggage stayed behind. CAD archive on NFS storage with offsite backup at a second provider (3-2-1 principle).
  • Cutover: on a Friday evening, aligned with production planning — Monday morning everything ran, and the workforce noticed nothing.

The stumbling block (honestly)

The CAD archive: 11 TB over the customer's uplink would have taken weeks. Solution: initial seed via encrypted disk shipped to the data centre, then delta syncs only. Unspectacular — but exactly the kind of detail that otherwise blows up projects.

−54 %running costs incl. care retainer
€8.5 → 3.9kmonthly cost (infra + ops)
0complaints from the workforce
RTO 4hdefined & confirmed in restore test

Afterwards

Hygge Care handles monitoring, patching, backups and the quarterly restore test — the IT lead receives a monthly report he forwards straight to management and the DPO. The data protection officer's question now has a documented answer.

"For the first time in years we feel like we own our IT instead of renting it."

Transparency: Both scenarios are anonymised, partly merged example journeys — figures and processes match typical projects of this scale. We're happy to name reference customers with contacts in a personal conversation, once their approval is in place. That, too, is part of "radically honest".

Your project could be the third scenario.

Tell us your starting point — in the intro call we'll tell you what your path would look like, what numbers to expect, and where the stumbling blocks are.

→ Book an intro call

30 minutes · no strings attached · we'll tell you if it's not worth it