Cloud exit
Cloud bills keep rising, managed services lock you in, and data-sovereignty rules keep getting stricter. We move your workloads to European or Canadian providers, bare metal or your own data centre, one at a time, and run them for you.
We reply within one business day.
Fixed scope
in writing before work starts
1 business day
reply time
At a glance
Compute and containers
Databases
Search and logs
Monitoring
Queues, cache and storage
In short
A cloud exit (or cloud repatriation) moves applications from hyperscalers such as AWS, Azure or Google Cloud to simpler, cheaper infrastructure: Hetzner, OVHcloud, other regional providers, colocation or your own servers. We replace managed services with open-source equivalents we already run every day, such as PostgreSQL or MySQL clusters instead of RDS, Azure Database or Cloud SQL, and Prometheus, Grafana and Zabbix instead of CloudWatch, Azure Monitor or Google Cloud Monitoring. Then we migrate workload by workload without a big-bang cut-over. Work starts with a fixed-price assessment that maps your bill, your dependencies and the expected savings.
Services
SQS, MSK, ElastiCache, Azure Service Bus, Event Hubs, Pub/Sub or Memorystore → RabbitMQ, Kafka and Redis/Valkey; S3, Azure Blob Storage or Cloud Storage → S3-compatible object storage.
CloudFront and Route 53, Azure Front Door and Azure DNS, or Cloud CDN and Cloud DNS → Cloudflare or Bunny; WAF, firewalls, VPN and secrets management.
Examples
Application, database and storage moved to EU servers, with the CDN and DNS switched last.
Databases moved with replication and a short switchover window.
Steady workloads moved out while serverless and spiky parts stay where they fit.
Search clusters self-managed and high-volume logs moved to ClickHouse.
How to start
Server costs at the new provider are paid directly by you. Many teams keep a few workloads in the hyperscaler where it still makes sense; we design for that.
1–2 weeks
Bill and workload analysis
Target architecture
Savings estimate and migration plan
Workload by workload
No big-bang cut-over
Rollback for every step
Fixed price per phase
24/7 monitoring and alerting
Patching and backups
Capacity and cost reviews
Response times in writing
Process
We read your cloud bill line by line, map dependencies and estimate the new monthly cost.
One non-critical workload moves first, so you see performance and operations before committing.
The rest moves in phases. We automate, monitor and operate the new platform.
Why now
For steady workloads, simpler providers and your own servers are often far cheaper, and they keep data in your jurisdiction.
Always-on servers, databases and storage cost much less on providers like Hetzner and OVHcloud.
Data transfer, NAT and managed-service premiums add up in ways that are hard to predict.
EU and Canadian providers keep data under local jurisdiction, which helps with GDPR and NIS2.
Proprietary managed services make price changes hard to walk away from.
With infrastructure as code and good monitoring, running your own stack is no longer the burden it was.
A hybrid approach keeps spiky or specialised workloads where they make sense.
An honest note before you book.
Highly variable or serverless workloads that benefit from per-second billing.
Heavy use of services with no practical open-source equivalent.
No one to operate the new platform, unless we run it for you.
DIY vs with us
FAQ
Straight answers. Ask us anything else on a call.
It depends on what you run. Savings are usually largest for steady workloads, databases, storage and data transfer, and smallest for spiky or heavily serverless workloads. The assessment gives you a line-by-line estimate before you commit.
Not if the replacements are run properly. We replace each managed service with a proven open-source setup, add monitoring and backups first, and move one workload at a time with a rollback plan.
Usually Hetzner or OVHcloud data centres in Germany, Finland, France or Canada, or your own servers. We pick the provider with you based on location, compliance and price.
No. A hybrid setup is often best: steady workloads move out, and a few services that really benefit from the hyperscaler stay.
Related
Next step
Send a few details, and pick up to 3 times if you would like a call. We reply within one business day.
You talk to the expert who would do the work
Written scope and price after the call
No obligation