Delivery, migration & support
Data infrastructure migration: cloud and on-premise
I move warehouses and pipelines between clouds, from on-premise to the cloud and from the cloud to on-premise — without data loss and with minimal downtime, verifying integrity at every step.
What is data infrastructure migration
Data infrastructure migration is moving a warehouse, ETL/ELT pipelines, integrations and dashboards to a new platform without data loss and with minimal downtime, while controlling integrity and the consistency of the numbers at every step. The relocation scenarios vary: between clouds, from on-premise to the cloud and from the cloud to on-premise, as well as a change of database. It's needed when changing providers, moving to regional or on-premise infrastructure, scaling under growing load, or when the platform has become expensive or outdated. I plan the move so that the business keeps running and the reporting reconciles after the switchover.
When you need a migration
What the service includes
Infrastructure audit and migration plan
I examine the current warehouse, pipelines, integrations and dashboards, assess the volume and risks, and put together a step-by-step relocation plan.
Data transfer with integrity control
I move historical data and increments to the target platform and reconcile volumes and control figures so that nothing is lost.
Pipeline migration and adaptation
I move the ETL/ELT pipelines to the new platform, rewrite the SQL and transformations, and set up orchestration on Apache Airflow and transformations with tests on dbt.
Dashboard and reporting migration
I switch dashboards and reports to the new warehouse, adapt the sources and queries, and verify that the visualizations work as before.
Number reconciliation
I run the old and new systems in parallel and compare data marts and key metrics — row by row and by aggregates, using SQL and Python. I make the numbers match before acceptance.
Cutover and support
I switch production over in an agreed window with minimal downtime, stabilize operations, and support the infrastructure after the move.
How the migration works
- 01
Audit and inventory
I examine the current infrastructure: warehouse, sources, pipelines, integrations and dashboards. I record volumes, dependencies and risks.
- 02
Designing the target architecture
I design the target data model and choose the migration strategy — phased or all at once, with the old and new systems running in parallel.
- 03
Migrating the schema and historical data
I move the warehouse structure and historical data to the target platform, controlling the integrity and completeness of the load.
- 04
Migrating pipelines and integrations
I move and adapt the ETL/ELT pipelines and source integrations, and set up scheduled loads and orchestration. For incremental synchronization and CDC during a hot migration, I use Kafka.
- 05
Reconciliation and validation
I compare data marts and key metrics on both platforms — row by row and by aggregates, using SQL and Python. I make the numbers match before acceptance.
- 06
Switchover and stabilization
I perform the cutover in an agreed window with minimal downtime, stabilize operations, and support the infrastructure after the move.
Projects
Analytics platform for the digital agency YA AGENCY
Analytical data warehouse and dashboards for Ash
End-to-end analytics and infrastructure migration for WiseAdvice-IT
Dashboards for the SaaS platform Entera
Analytical data warehouse for a chain of bars and restaurants
Analytical data warehouse for the Neolombard pawnshop chain
Unified data warehouse and dashboards for the delivery service Edostav
Data warehouse for the financial analytics project VISUALSECTORS
Tech stack
dbt 
OpenMetadata Clients
Testimonials
Migration isn’t just copying data to a new platform — it’s moving the whole chain from sources to dashboards so the business never stops. I carry out the move with parallel runs and number reconciliation, so that after the switchover the reporting matches the old one and the team keeps working without disruption.
Shall we discuss your task?
FAQ
How do you avoid business downtime during a migration? +
I run the old and new systems in parallel and keep the data synchronized incrementally, while performing the final switchover in an agreed window. This way the business keeps running on the current platform until the new one is ready for acceptance.
How do you make sure no data is lost and the numbers match? +
I reconcile volumes and data marts on both platforms — comparing rows and aggregates with SQL and Python, checking key metrics and discrepancies before the switchover. Acceptance happens only once the numbers match.
Can you migrate between different databases, for example BigQuery to ClickHouse? +
Yes. A change of database is a common scenario: I move the schema and data, rewrite the SQL and adapt the pipelines to the specifics of the target platform, then reconcile the result against the source system.
What happens to the dashboards and integrations? +
I move and adapt the dashboards and integrations to the new warehouse: switch the sources, rewrite the queries and verify that the reports compute the same numbers as before.
Does migration support on-premise or regional infrastructure? +
Yes. I migrate infrastructure to on-premise or a regional cloud, including a change of database and cloud provider, with the same approach — data reconciliation and minimal downtime.
How long does a migration take? +
I don't quote fixed timelines in advance: they depend on the data volume, the number of pipelines and integrations, and the chosen strategy. I give an estimate after auditing the current infrastructure.