Platform & governance

Platform support levels

Development documentation. This guide follows a reviewed snapshot of main and may describe capabilities not yet in Stable. Check Preview labels and release notes. Some executable names and paths remain unchanged for compatibility. Legacy names explained ↗

KratosForge works with Microsoft Fabric, Azure Databricks and Snowflake, and runs work on this PC. This page says how far each one is supported, and what evidence that rests on. It is generated from the code (fabric_desktop/platforms.py), and a test keeps it equal to the code, so it can't claim more than the app does.

Levels#

Level Name What works
0 Discovered Known; nothing integrated yet.
1 Connected Sign in, and list what the platform holds.
2 Development Level 1, and run work.
3 Integrated Level 2, and compare or validate results, and publish or deploy.
4 Optimized Level 3, and measure usage and cost.
5 Enterprise ready Level 4, and security, reliability and lifecycle checked for an enterprise rollout.

Each level is worked out three times, from three strengths of evidence:

  • Built: the code does it, with automated tests in this repository. Remote platforms are tested against fakes of their APIs.
  • In use: it has been used against the real platform in normal work, without a recorded test.
  • Verified: a recorded test against the real platform passed. On this PC the tests run the real engines, so tests count as verified there.

Planned work never counts toward a level.

Where each platform stands#

Manifest 1.

Platform Built In use Verified To verify the next level
Microsoft Fabric 4 Optimized 2 Development 0 Discovered sign in, list what the platform holds
Azure Databricks 2 Development 0 Discovered 0 Discovered sign in, list what the platform holds
Snowflake 0 Discovered 0 Discovered 0 Discovered sign in, list what the platform holds
This PC 4 Optimized 4 Optimized 4 Optimized checked for an enterprise rollout

Nothing remote is verified yet: the Fabric live test (fabric-desktop live-test), the tenant smoke test and the Databricks live test are built, and each run needs a test workspace and the owner's go. Snowflake has connection profiles and Snowflake SQL tried on this PC (see Engines), but no sign-in yet, so it stays at level 0.

What each level rests on#

platforms.manifest("fabric") (and "databricks", "snowflake", "local") lists, for each area, the evidence and where it comes from: an engine's own capability claims (see Engines), or a feature with the modules that do it and the tests behind it. The manifest also lists each platform's sign-in methods, engines, SQL targets, notebook formats and, for Fabric, how many pipeline activity types run on this PC.

Need a hand?

Find a fix in troubleshooting, or share a reproducible issue with the community.

Get support ↗Improve this guide on GitHub ↗

Search documentation

Start typing to search all 172 guides.

Search runs in your browser esc to close