The engineering decision
Repository data is only part of a migration. Plan identities, permissions, pipelines, integrations and operating ownership alongside the transfer itself.
On this page
Agree what a successful move includes
Create an inventory of repositories, large files, issues and reviews, users, permissions, CI configuration, credentials, registries and integrations. For each item, record its destination, migration method, validation check and owner. Mark exclusions explicitly.
GitLab’s import matrix differs by source platform and method. Direct transfer has documented migrated and excluded items; a Git repository transfer is not a guarantee that every associated service or setting moves with it.
Rehearse a representative slice
Choose projects that expose the difficult parts of the estate, not only the smallest repository. Include the identity mapping, a real pipeline, relevant large files and an external integration. Ask the engineers who use those projects to inspect the result.
For example, GitLab’s GitHub importer has optional content and prerequisites that affect the result. Use the documentation for the selected importer and version when deciding what to include. Compare the migrated artefacts with the agreed inventory rather than relying only on a success status.
Agree the cutover decision
- Define the write-freeze or final-sync approach and communicate which system is authoritative.
- Agree who can accept the result and which validation failures stop the cutover.
- Document how users, CI jobs and integrations find the new destination.
- Rehearse recovery, including the data changes that could occur after switching.
Availability, migration windows and recovery time depend on the source, method, volume and dependencies. Commit to a measured plan and tested checkpoints instead of a blanket zero-downtime promise.
Validate the workflow that matters
Check representative repository history, membership/access, issue/review relationships, pipeline execution, artifacts and integration behaviour. Track exceptions with an owner and acceptance decision. Security obligations should be mapped to the organisation’s requirements and evidence, not described as automatically satisfied by moving tools.
Agree who owns support during the transition and how the source will be retained, restricted or decommissioned after acceptance.
Leave the team ready to operate
A useful handover includes operating notes, access ownership, known exceptions, monitoring and a practical walkthrough. Training should use the migrated projects so engineers learn the workflow they will actually follow.
OTTRA can scope the migration work, support team enablement and discuss continuing operation through Cloud Choice™. Share the inventory and constraints; scope, timing and cost are agreed before delivery.
Primary sources
Source scope and product terms can change. Checked on 7 November 2026.





