Delbridge Solutions
Delbridge Solutions

Enabling a Zero-Downtime Database Migration for a Major Retail Group

Talk to an Expert

Modernizing Cloud Database Infrastructure

For one of South America’s largest multi-format retail groups, operating across grocery, department stores, home improvement, and financial services, database performance and reliability are central to keeping customer-facing systems and internal operations running smoothly.

 

To modernize its cloud database infrastructure, the retailer engaged Delbridge to migrate a critical workload from Azure CosmosDB (Mongo API) to MongoDB Atlas. The objective was a clean, fully validated migration across both Dev and Production environments, executed without operational disruption and with a clear handover that allowed the client’s internal team to control the final cutover on their own timeline.

 

The Delbridge Approach

Delbridge delivered the migration as a structured, end-to-end engagement spanning roughly 20 working days. The scope covered approximately 27 GB of data across 8 collections, with the source schema preserved as-is so that application and microservice changes could remain entirely on the client’s side. The final cutover was likewise kept under client control, with Delbridge providing the tooling, validation, and documentation needed to execute it safely.

 

Rather than treating the migration as a single high-risk event, Delbridge structured the work around repeatable dry runs, automated validation, and continuous data replication. This approach allowed the client’s team to gain confidence in the process before committing to production, and to schedule the cutover at a time that suited their business operations.

Hands-On, End-to-End
MongoDB Atlas clusters were provisioned and configured for both Dev and Production environments, with performance baselines established to confirm parity with the existing CosmosDB workload. DSync was deployed as a single data movement layer covering both the initial bulk migration and ongoing change data capture, removing the need for separate tooling between phases. A custom document-comparison validation harness was built in Node.js, packaged for Docker, and driven by configuration files referencing each collection and its document IDs. This harness verified field-level parity between source and target after every migration pass.
Long-Term Resilience
Three full dry runs were executed across Dev and Production environments to de-risk the production migration, with the validation harness run after each pass to confirm data integrity. Both environments were performance-tested against baseline expectations before any production activity took place. Post-migration monitoring was set up on Atlas to give the client clear visibility into the new environment, and a zero-downtime cutover playbook was prepared and handed over to the internal dev team so they could execute the final switch at their own discretion, supported by live CDC.

The Challenge: Migrating a Production Workload Without Disruption

Moving a live workload between two different cloud database platforms required careful planning around data integrity, replication continuity, and operational risk. The client was not looking for a basic lift-and-shift, but for a migration that could be validated rigorously and handed over in a way that preserved their control over the final cutover.

A cross-platform migration with no room for data loss
The move from Azure CosmosDB (Mongo API) to MongoDB Atlas spanned approximately 27 GB across 8 collections, and required absolute confidence that every document arrived intact.
A need for both bulk migration and ongoing replication
The migration could not rely on a one-time data copy. Continuous change data capture was needed to keep source and target aligned through the client-controlled cutover window.
Field-level validation requirements
Confirming row counts was not sufficient. The client needed verifiable parity at the field level across all collections before signing off on production.
Multi-environment coordination
Work had to be carried out across both Dev and Production environments, with each migration pass repeatable and consistent.
A client-controlled cutover boundary
Application and microservice changes remained on the client's side, as did the final cutover itself. Delbridge needed to deliver a process and toolkit that the internal team could operate independently.
Performance parity expectations
The new Atlas environment had to meet or exceed the performance baseline established under CosmosDB, with no degradation in day-to-day responsiveness.

Delivery and Execution

A key part of the work involved provisioning and configuring the new MongoDB Atlas clusters for both Dev and Production, and standing up DSync as the data movement layer covering both the initial bulk migration and ongoing change data capture. Using a single tool across both phases reduced operational complexity and gave the client a consistent replication path through to cutover.

Delbridge also built a custom document-comparison validation harness in Node.js, packaged to run in Docker and driven by configuration files specifying collections and document IDs. This harness was used after every migration pass to confirm field-level parity between CosmosDB and Atlas, giving the client objective evidence of data integrity rather than relying on row counts or sampling.

Three full dry runs were executed across Dev and Production before the production migration itself, with the validation harness run at each stage. Both environments were performance-tested against the CosmosDB baseline, and post-migration monitoring was configured on Atlas to support ongoing visibility.

Taken together, these activities allowed the production migration to be executed cleanly, with live CDC keeping the source and target in sync. The client’s internal team retained full control over the final cutover, supported by a zero-downtime playbook that removed the need for any data-freeze window.

Results

The engagement gave the retailer a fully validated migration path from Azure CosmosDB to MongoDB Atlas, executed end-to-end in roughly 20 working days. By combining a single replication layer for bulk and CDC, a custom validation harness, and multiple dry runs, Delbridge helped create a migration process that was both rigorous and repeatable.

 

Just as importantly, the engagement was structured to respect the client’s operational boundaries. Application changes and the final cutover stayed on the internal team’s side, supported by a clear handover playbook that allowed them to execute the switch at their own pace without a data-freeze window.

 

For a retail group of this scale, this kind of migration extends beyond moving data between platforms. It helps reduce delivery risk, preserve operational control, and give internal teams the confidence to take ownership of the new environment from day one.

  • Migration of ~27 GB across 8 collections completed with no incidents
  • 100% data integrity confirmed via a custom field-level validation harness
  • Three successful dry runs executed before the production migration
  • Live CDC maintained between CosmosDB and Atlas through the cutover window
  • Atlas performance met or exceeded the CosmosDB baseline
  • Post-migration monitoring configured on Atlas
  • Zero-downtime cutover playbook delivered to the internal dev team

Whether you are modernizing legacy systems, migrating large and complex datasets, transforming data for MongoDB, or replacing outdated migration methods, Delbridge delivers a structured, scalable migration framework built to move your organization forward with speed, control, and confidence.

Learn More

What do we deliver ?

Create your engagement
“Always-on” Monitoring
1
A dedicated watch of your running platform to ensure health, operation, status, and configuration for optimal use.
Timely Incident Response
2
An eye on platform health and management of issues if and when they occur, aligned to service levels important to you.
Cluster Provisioning
3
A proactive involvement to ensure that the cluster is operating as expected, and appropriately sized.
Backup and Recovery Management
4
Formulation and execution of backup procedures to minimize the impacts of an unforeseen outage.
Upgrades and Change Management
5
Support the rollout of platform changes including version upgrades and your application deployments.
Risk Mitigation
6
Continuous exploration of optimizations that can be applied to reduce operational risk of your platform now and into the future.

Project Details

Large Retail Group
South America
MongoDB
Adiom provides an out-of-the-box connector for migrating from Azure CosmosDB to MongoDB, which removed the need for custom replication tooling and gave the engagement a proven, purpose-built path between the two platforms from day one.

Project Stats

27 GB
Data migrated
8
Collections migrated
3
Full dry runs executed
20 Working Days
Total engagement duration
100%
Data integrity verified
Zero
Downtime at cutover
Live CDC
Maintained through handover
Our MongoDB Expertise

Our MongoDB Expertise

14+
Years of Experience
100+
Successful MongoDB Migrations
45+
MongoDB Specialists
Our MongoDB Expertise

Fast Start

Migration foundation deployed with documented patterns and standards.​Environment-ready and built to scale.​​

Best Suited For: ​

  • Small budgets​
  • Fast turnarounds that demonstrate value​
  • Proving capability​
  • Building momentum ​

White Glove

Fully-managed delivery process where we deliver strategy, execution, and promotion along with app changes to support MongoDB

Best Suited For:​

  • Highly occupied teams​
  • De-risking delivery​
  • MongoDB novices ​
  • Hitting deadlines​​