One of the largest retail and financial services business ecosystems in Latin America faced the massive challenge of migrating its large-scale development infrastructure from a decentralized, non-standardized GitLab instance to GitHub Enterprise Cloud. With more than 21,000 repositories distributed across 17 different business platforms, the existing situation was characterized by high structural fragmentation, a lack of documentation regarding code ownership, and significant technical debt in legacy CI/CD pipelines.

To reduce the risks of this transformation, CleverIT designed and coordinated a comprehensive, data-driven strategy that included an advanced technical assessment, architectural classification, and a phased migration framework. With the support of GitHub and Microsoft, CleverIT successfully cataloged the entire ecosystem, mapped 224.4 GB of source data, and classified over 16,000 active pipelines, including thousands of highly complex, mission-critical pipelines with up to 32,000 lines of expanded code.

This work establishes a secure, governed, and highly automatable roadmap to consolidate engineering operations onto a top-tier platform, accelerating time-to-market for solutions, strengthening source code governance, and improving digital experiences for millions of retail and banking customers across the region.

‍

A large-scale infrastructure with complexity and governance challenges

The scale and operational distribution of our client's engineering infrastructure created significant technological, organizational, and operational challenges that threatened business agility and system stability.

‍

• Significant CI/CD technical debt: Of the more than 21,000 repositories, 16,738 (79.1%) had active CI/CD pipelines that required a complete architectural conversion to GitHub Actions.
  ‍
• Extreme pipeline complexity: More than 6,389 pipelines presented critical complexity, exceeding 1,000 lines of expanded code due to the use of nested templates, external integrations, and extensive conditional logic. Some exceptional cases reached 32,601 expanded lines and accumulated 552,015 historical executions, which required an exhaustive structural redesign.
  ‍
• Organizational fragmentation: Engineering operations were distributed across 17 different business units and platforms. Nearly 49% of the repositories lacked clear classification or defined owners—identified as "NO FAI"—which made it difficult to align namespaces and securely assign migration responsibilities.
  ‍
• High operational risk: The heterogeneity of the repositories and the large volume of artifacts, including 2,483 repositories larger than 100 MB, increased the risk of production outages, made resource estimation difficult, and could have caused serious business impact if migrated directly without prior adaptation.

‍

A migration strategy based on data, complexity, and automation

CleverIT designed a meticulous, phased migration and modernization strategy to safely transition the client to GitHub Enterprise Cloud, while simultaneously modernizing their core DevOps practices.
‍

• Data-driven discovery and assessment: Automated scripts were implemented to audit, map, and inventory the entire GitLab instance, providing visibility into code volume, Large File Storage (LFS) extensions, and structural patterns.
• Complexity classification matrix: A tiered categorization model was developed to group pipelines based on their number of lines of code and dependencies, allowing for specific architectural interventions to be defined prior to migration.
• Phased migration framework (wave model): A progressive migration plan was structured using automated GitHub tools, isolated migration testing, parallel validation flows, and rollback plans to minimize operational disruptions.
• Modernization via GitHub Actions: Reusable and standardized GitHub Actions templates were designed to replace highly fragmented legacy configurations, incorporating modern secret management and environment governance.
• Strategic collaboration: A strong partnership between vendors was leveraged. GitHub provided direct guidance from product specialists, while Microsoft validated CleverIT's technical expertise, strengthening client confidence in project execution.

‍

From technical assessment to a 12-month roadmap

The project was executed by a specialized CleverIT team, consisting of 1 Project Manager, 1 senior DevOps and GitHub technical lead, and 5 engineers specializing in DevOps and GitHub.

The work was organized through a 12-month roadmap: 2 months dedicated to a comprehensive assessment and 10 months for the phased migration execution. Azure DevOps was used to track milestones, blockers, and scope changes on a daily basis.

‍

The implementation included four main architectural milestones:

1. Current state assessment: Comprehensive analysis of the 21,165 repositories, identifying size limits, LFS requirements—with 2.8 GB of data identified—and pipeline volume.
2. Migration planning and target structure design: Collaborative design of the new organizational hierarchy, portfolio standardization, and definition of repository namespaces in GitHub Enterprise Cloud, in coordination with the client's management and technical teams.
3. Migration testing (dry runs): Simulation of automated phased migrations for critical repositories with multiple providers to validate data integrity, permission mapping, and performance parameters.
4. Execution of automated migration waves: Migration of production environments in batches planned according to business priorities and the complexity of each unit, avoiding risks associated with single points of failure.

‍

Over 21,000 repositories analyzed for a secure and scalable migration

The structured approach provided comprehensive visibility and a solid roadmap for our client's enterprise DevOps transformation.

Instead of opting for a direct migration without a prior strategy, the initiative allowed for the inventorying, analysis, and characterization of 21,165 repositories, resolving the lack of consolidated visibility and establishing a clear view of the 224.4 GB of source code evaluated, including 2.8 GB of assets stored using Git LFS.

To reduce operational risks and address the hidden complexity in custom templates, we identified 16,738 repositories with active CI/CD that required their pipelines to be converted to GitHub Actions.

The assessment also addressed significant gaps in governance and accountability: 100% of repositories without proper tracking were identified and set aside for immediate reclassification. This allowed us to tackle the issue of repositories that previously accounted for nearly 49% of the total and were classified as "NO FAI," with no clearly defined owners.

Finally, the pipeline engineering analysis resulted in a phased migration plan, based on dependency mapping and a scalable effort estimation model.

‍

The ecosystem classification was distributed as follows:

• 5,700 repositories without CI/CD.
• 1,218 pipelines of simple complexity.
• 3,417 pipelines of moderate complexity.
• 2,412 pipelines of high complexity.
• 1,749 pipelines of very high complexity.
  ‍

Additionally, 6,389 pipelines of critical complexitywith over 1,000 lines of expanded code, were specifically identified for an architectural redesign to enable a seamless and scalable modernization.

‍

A more secure platform to drive digital innovation

By establishing a unified, top-tier engineering standard on GitHub Enterprise Cloud, the client is moving toward reduced technological risk, stronger source code governance, and more efficient global collaboration across its 17 lines of business.

For end users, this transformation helps create a more secure digital environment. Automated workflows and advanced security capabilities help protect the personal and transactional data of millions of customers, fostering the development of faster, more robust, and more resilient software solutions for retail and banking.

‍

Three key lessons for transforming complex technology ecosystems

• The pre-migration assessment is fundamental: For projects involving more than 20,000 repositories, performing a direct migration without analyzing expanded pipeline code can lead to conversion failures from the very early stages. Categorizing repositories by complexity allows for more accurate planning of resources and timelines.
  ‍
• Standardization is better than replication: Highly complex pipelines, which can reach up to 32,000 lines of code, are often the result of legacy templates and poor management. Migration represents an opportunity to implement reusable and efficient GitHub Actions components, significantly reducing future technical debt.
  ‍
• Reorganizing governance is a prerequisite: Normalizing portfolios, namespaces, and ownership assignments before moving files prevents replicating historical disorganization in a clean GitHub Enterprise Cloud instance.

‍

From planning to standardized operations on GitHub

• Scaling migration waves: Systematically execute the automated migration of remaining repositories to GitHub Enterprise Cloud, following the prioritized and validated roadmap.
• Phasing out legacy infrastructure: Gradually decommission on-premises or cloud GitLab instances as business platforms complete their validation periods on GitHub.
• Implementing global GitHub Actions standards: Deploy reusable workflows validated by CleverIT across all 17 business units, preventing configuration drift and ensuring unified compliance criteria for CI/CD.
• Driving enterprise-wide adoption: Launch training and adoption programs to strengthen the client's engineering team competencies in the native GitHub ecosystem, advanced environment controls, and integrated secret management.