Who Uses Flux and Why It Scales

▶ Watch (0:13)

Flux reconciles cluster state from Git. That model lets teams manage hundreds of clusters declaratively without manual operations. Azure ships a Flux extension in its marketplace. AWS also trusts it. In the telco space, multiple carriers built Silva, an open-source project on top of Flux that provisions a management cluster and handles workload cluster lifecycle. Large financial institutions adopt Flux specifically for scale, because manual operations do not scale past a certain point.

Helm Support and Server-Side Apply

▶ Watch (0:48)

Flux has a deep integration with the Helm SDK. Support for Helm v4, released late last year, shipped in Flux after close collaboration with the Helm project. Drift detection works well given how Helm was designed internally. Flux also uses server-side apply natively, a Kubernetes feature that prevents conflicts when multiple tools manage the same cluster resources. Without it, competing controllers frequently corrupt each other’s state.

Health Checks and CEL Expressions

▶ Watch (3:36)

Bootstrapping a cluster requires steps in a specific order. Flux handles this through dependency declarations between Kustomizations and Helm Releases, combined with health checks. A recently introduced feature adds CEL expression support. Different Kubernetes controllers express resource health differently in the status field. CEL expressions let operators write a custom health condition per resource kind, so the GitOps pipeline only advances once the right signal appears in the right field.

Flux Operator, UI, and MCP Server

▶ Watch (5:08)

Control Plane released the open-source Flux Operator, which lets teams manage Flux itself declaratively. The operator also ships a UI, a feature users requested for a long time that has seen strong adoption since release. It now includes an MCP server. Pimenta noted that vibe coding is inevitable, so the project released agent skills as well. OCI artifact support ships configuration through container registries, reusing the security tooling already built around image distribution.

OpenTelemetry Traces: What Is Shipping Next

▶ Watch (6:25)

Flux recently released an OpenTelemetry integration that produces traces for Flux resources. The next increment will add a deep trace tree with proper correlation across resources. That means operators will be able to follow a reconciliation event through every Flux object involved, not just see isolated spans per resource. Pimenta closed by pointing attendees to a QR code for further conversation.

Notable Quotes

you can declaratively manage the tool that manages things for you in a declarative way Matheus Pimenta · ▶ 05:21

VIP coding is inevitable. So uh we obviously uh following the trends Matheus Pimenta · ▶ 05:46

manual operations don’t scale right and flux makes it possible for you to manage uh hundreds of clusters declaratively via GitHubs Matheus Pimenta · ▶ 02:05

Key Takeaways

  • Flux manages hundreds of clusters declaratively; Azure ships it as a marketplace extension.
  • CEL expressions let operators define custom health conditions per resource kind in GitOps pipelines.
  • The Flux Operator adds a UI, MCP server, and agent skills, all released recently.