Questions for considerations when discussing importer framework

Everyone can contribute. Help move this issue forward while earning points, leveling up and collecting rewards.

• Close this issue

There's a proposal for imported framework in issue https://gitlab.com/gitlab-com/Product/-/issues/12908/ that states what framework would enable and include:

• Unify the experience of using the importers, ensuring consistent user experience. Quotes from proposal issue:
• • "Framework includes characteristics of importers that are common across. For e.g. Error handling, Evaluate tool, Progress tracker etc."
• • "Framework shows the evaluation of the import in preparation for the actual import."
• • "Framework offers one entry point to start migration from any source (GitHub, Bitbucket, GH Actions)."
• • "Framework shows progress of the import."
• • "Framework offers re-try logic ( To retry importing only failed parts.)"
• • "Framework shows the output with success / error responses to the users."
• And make implementing new importers easier and faster.

Questions

• What in the backend can be abstracted out into the framework?
• What in the frontend can be abstracted out into the framework?
• What would need to be build from scratch?
• What percentage of an importer would such framework substitute?
• How big of an backend effort to abstracted what's already there out into the framework could that be? For the team consisting of 3 BE, would it take ~2-3 months, ~6 months or ~1 year, more?
• How big of an frontend effort to abstracted what's already there into the framework could that be? For 2 FE, would it take ~2-3 months, ~6 months or ~1 year, more?
• The assumption is that the framework would substantially reduce the effort that other product groups would need to put in when implementing new importers. Please validate the assumption.
• The assumption is that the framework would support use cases like importing GitHub actions (different than the project with repo and issues etc.) Please validate the assumption.
• Would it be advisable for the technical standpoint to port existing importers to this framework?
• If yes, how big of an effort could that be? For the team consisting of 3 BE and 2 FE, would it take ~2-3 months, ~6 months or ~1 year, more?