<aside> 👐 Maintained ****by @Christian Weichel . Contribute to this page using following our contributor flow.

</aside>

This page outlines architectural goals for the near future. Considering the ambitious Product Roadmap, we'll have to align closely and try and implement those goals as "side-effects" of that roadmap.

Compared to the product roadmap, this roadmap outlines goals and work that is not immediately user facing, but affects development speed, reliability, TCO and operations.

Roadmap

• 2021 Q1
  • Reduce Variance: Registry Facade
  • Reduce Variance: User-Namespaced Workspaces
  • Basic Requirements: User-Namespaced Workspaces
• 2021 Q2
  • Shorter Paths: Headless Workspace Logs
  • Reduce Variance: Full Workspace Backup
  • Basic Requirements: Full Workspace Backup

Technical Debt

In the first quarter/half of 2021 we'll try and catch up on variance and technical debt we've introduced over the last two years. This will position us better for the features that are to come.

Shorter Paths

Pathways in Gitpod have become long. Headless workspace logs (prebuild logs) are a good example: they travel far before reaching the user's browser.

But also operations from within a workspace can be far traveled, e.g. port exposure:

These long pathways make the system

• brittle, because at each interface there's a chance for something to break
• complex, because they it hard to follow things throughout the system
• costly, because all those intermediary steps cost CPU and RAM.

Some of those pathways are there because that was the first solution we came up with (e.g. prebuild logs), others because of prior design principles (e.g. user auth in server). We would do well with shortening those paths; although not at any cost.

Examples and known issues