kubernetes

The CNCF Technical Oversight Committee has voted to accept Tekton, the Kubernetes-native CI/CD framework with 600-plus contributors and 11,000 GitHub stars, ending its tenure as a CD Foundation graduated project.

IBM Research, Red Hat, and Google Cloud donated llm-d, a Kubernetes-native LLM inference framework, to the CNCF as a sandbox project at KubeCon Europe 2026, backed by NVIDIA, CoreWeave, AMD, and other partners.

NVIDIA transferred its Dynamic Resource Allocation driver for GPUs to community ownership under the Kubernetes project at KubeCon Europe 2026, while Google open-sourced the GKE Cluster Autoscaler and Broadcom donated Velero to the CNCF Sandbox.

Cilium 1.19 introduces beta ztunnel integration for transparent mTLS encryption without sidecars, strict encryption modes for WireGuard and IPsec, and stable Multi Pool IPAM, as the CNCF-graduated eBPF networking project marks a decade of development.

Google Cloud now accepts OTLP-format metrics in Cloud Monitoring and offers a fully managed OpenTelemetry Collector pipeline for GKE, joining a broader industry shift toward vendor-neutral observability driven by open standards.

The next Kubernetes release locked its feature set on March 18, delivering 80 enhancements that graduate User Namespaces and four Dynamic Resource Allocation features to GA while introducing gang scheduling and workload-aware placement for AI training clusters.

The Cloud Native Computing Foundation has graduated Dragonfly to its highest maturity level, recognizing the peer-to-peer distribution system's role in scaling container and AI workloads across production environments at Alibaba, ByteDance, Datadog, and Intel.

The largest cloud native conference opens March 23 in Amsterdam, introducing Agentics Day alongside 224 sessions spanning AI workloads, platform engineering, and security.

The community-maintained Ingress NGINX controller goes read-only in March 2026. With no future security patches and four new CVEs disclosed in February, organizations running roughly half of all Kubernetes ingress traffic must now choose between Gateway API and alternative controllers.

Kubernetes 1.35 adds in-place pod-wide restarts and other scheduling and scaling updates as operators report broader production adoption for AI workloads.