AI
How the CNCF Maturity Ladder Works in Practice
CNCF hosts more than 230 projects spread across three maturity levels: sandbox, incubating, and graduated. Sandbox projects target small, defined problem spaces. Incubating projects show production adoption and multi-organization contributions, which protects vendor neutrality. Graduated projects are stable and established. A fourth state, archival, gets less attention but matters. Projects that stall on contributions should take their lessons and redirect energy toward existing initiatives or new open source work rather than linger.
The TOC Due Diligence Process
When the TOC evaluates a project for maturity advancement, it checks governance, contributor strategy, security practices, and engineering principles. Governance must be defined and applied, not just documented. Security review confirms that vulnerabilities can be reported and addressed on time. The TOC also uses LFX Insights as a reference point. For Carbon Cutter, LFX data showed room to improve on organizational and contributor dependency, which prompted a recommendation to form a steering committee drawn from end users.
The TAB as the Connector Between Users and Projects
The Technical Advisory Body collects production feedback from adopters and routes it back to the TOC before maturity decisions are made. Ahmed Bebars described three input channels: LFX Insights adoption data, reference architectures built by other end users, and adopter interviews. The interviews are unfiltered accounts of what happened in production. That ground-level detail complements the TOC’s technical checklist and ensures maturity assessments reflect real deployment conditions, not just code quality metrics.
From GitHub Issue to Merged Pull Request
Chad Beaudin’s fictional company Cycle AI identified a gap in Carbon Cutter’s observability support and filed a GitHub issue. The project maintainer confirmed the gap was real, found that multiple adopters had commented on the same issue, reviewed his pull request, and approved it on stage. The feature shipped to alpha first, with a test suite available for other end users to validate against their own use cases. A GA milestone went onto the road map for the next major release.
Closing the Loop: Contributing a Reference Architecture Back
After running Carbon Cutter in production for several months, Cycle AI built its own reference architecture and wanted to share it. The process is a GitHub issue against the TAB repo. Fill in the organization name, the team, and the projects used. The TAB connects back, reviews the submission, and publishes it so other end users can build from proven work. Chad’s arc ended with his architecture going live, completing one full turn of the feedback loop from consumer to contributor.
Notable Quotes
project health is not just a dashboard. It’s actually a real problems that you have seen in production. Ahmed Bebars · ▶ 06:53
These are the real candid unvarnished conversation on like what exactly happened in production. Ahmed Bebars · ▶ 07:20
it’s great to put a face to your GitHub handle. Karena Angell · ▶ 08:11
don’t just consume cloud native and contribute to it and close the loop. Katie Gamanji · ▶ 12:55
Key Takeaways
- CNCF’s 230-plus projects sit across sandbox, incubating, graduated, and archival states, each with distinct production implications.
- TOC due diligence checks governance, contributor diversity, security response, and a documented road map with milestones.
- End users can close the loop by filing a TAB GitHub issue with their reference architecture for other adopters to reuse.
About the Speaker(s)
Katie Gamanji is a Senior Field Engineer at Apple and a CNCF TOC member. She previously built platforms at Conde Nast and American Express, led the CNCF End User Community, authored the Cloud Native Fundamentals course, and serves as lead for the CNCF KCNA certification. She is a #WITAwards and #TechWomen100 winner and advocates for women in STEM.
Ahmed Bebars is a Principal Engineer on the Developer Platforms mission at The New York Times, where he specializes in cloud infrastructure and Kubernetes-based solutions, with a focus on secure runtime environments for engineering teams.
Chad Beaudin is Chief Engineer and Technical Fellow at the Boeing Software Factory, where he drives the technical vision for cloud native software development across thousands of engineers in regulated environments.