Flux is a set of continuous and progressive delivery solutions for Kubernetes, and they are open and extensible. The APIs of Flux are stable now.
Employees
1-10
AI Classification
SaaSInfrastructure/DevTools
Flux CD is a CNCF Graduated open-source GitOps continuous delivery tool for Kubernetes that automatically synchronizes clusters with Git repositories, supporting GitHub, GitLab, Bitbucket, S3, and major container registries with full OCI support. It is used at enterprise scale by Deutsche Telekom, Cisco, Microsoft Azure, and AWS, with commercial support provided by ecosystem partners including ControlPlane, Fairwinds, and GitLab.
Deep Intelligence
Buying Signals
Flux v2.8.0 released February 2026 with Helm v4 support, server-side apply capabilities, and enhanced health checking — active release cadence signals continued investment
Flux MCP Server unveiled at KubeCon 2025, connecting AI assistants directly to Kubernetes clusters — strategic expansion into AI-driven automation use cases
New Flux Web UI introduced at KubeCon 2025 — signals product maturity push toward broader, less CLI-dependent adoption
FluxCon held at KubeCon North America in Atlanta (November 2025) — first dedicated community conference indicating growing ecosystem and commercial interest
CNCF assumed formal stewardship after Weaveworks closure (February 2026) and announced funding for mentorships and security work — institutional continuity confirmed
Account Scoring
Tier BLow Churn RiskProduct Launch
Pain Signals
Weaveworks, the original commercial backer, ceased operations in February 2026, creating potential support fragmentation and uncertainty for enterprise adopters evaluating long-term vendor risk
Commercial support is distributed across a coalition of smaller partners (ControlPlane, Aenix, Aviator, Fairwinds, Gimlet) with no single primary commercial vendor — fragmented support model may be a barrier for risk-averse enterprises
Core maintainer team appears small relative to the scale of enterprise adoption (Deutsche Telekom managing 200+ clusters with 10 engineers), suggesting limited dedicated engineering resources