What’s New in Cloud Native: Helm 3 Beta, Cilium 1.6, Kubernetes Enhancements & Top Open‑Source Projects
This week’s cloud‑native roundup covers the Helm 3 beta release, Cilium 1.6’s full kube‑proxy replacement, new Kubernetes apiserver request‑priority KEP, HA master StorageVersion API, scheduling framework redesign, etcd compact‑batch tweaks, knative performance work, plus recommended open‑source tools and reading material.
Industry News
Helm 3 first beta (v3.0.0‑beta.1) released – focuses on final refactoring and back‑porting Helm 2 features. Related link: https://github.com/helm/helm/releases
Cilium 1.6 released – now able to replace kube‑proxy 100% and includes performance test results. Related link: https://cilium.io/blog/2019/08/20/cilium-16/
Pivotal’s open‑source controller kpack – provides a resource controller for building and updating container images. Related link: https://github.com/pivotal/kpack
Upstream Important Progress
Kubernetes apiserver request prioritization (KEP) – introduces finer‑grained classification of observed requests and assigns each request to a concurrency pool based on priority. The existing max-in-flight limits mutating and readonly requests, but this KEP adds priority‑aware limits for other request types. References: https://github.com/kubernetes/enhancements/pull/1218, https://github.com/kubernetes/enhancements/blob/master/keps/sig-api-machinery/20190228-priority-and-fairness.md
HA master StorageVersion API – detects inconsistent storage versions across apiserver instances during rolling upgrades and blocks migrator progress when versions differ. Reference: https://github.com/kubernetes/enhancements/pull/1220
Scheduling framework redesign – a plugin‑based architecture for the Kubernetes scheduler, allowing new scheduling features to be added as plugins. Expected to reach beta in Kubernetes 1.17. References: https://kubernetes.io/docs/concepts/configuration/scheduling-framework/, https://github.com/kubernetes/enhancements/issues/624
Other updates
Fix for kubectl -f on Windows (explicit follow symlink). https://github.com/kubernetes/kubernetes/pull/81747
CustomResourceDefaulting promoted to Beta and enabled by default; v1beta1 webhooks/CRD types deprecated. https://github.com/kubernetes/kubernetes/pull/81872, https://github.com/kubernetes/kubernetes/pull/81899
Allow apiserver to start with HTTP/1 only (DisableHTTP2). https://github.com/kubernetes/kubernetes/pull/81860/files
Scale client now supports non‑namespace (cluster‑wide) resources. https://github.com/kubernetes/kubernetes/pull/81507
Kubelet now accounts for pod‑overhead when calculating pod resource usage. https://github.com/kubernetes/kubernetes/pull/79247
etcd Project Updates
MVCC: default compact batch size set to 1000 and interval to 10 ms. Adjusted per‑cluster based on QPS testing.
Raft: allow learner nodes to vote in special cases, fixing a scenario where a promoted learner could not obtain votes after a voter failure. https://github.com/etcd-io/etcd/pull/11034, https://github.com/etcd-io/etcd/pull/10998
Knative Project Updates
Serving and Eventing have entered a performance‑optimization phase with benchmarks (deployment benchmark, activator throttler toggling, eventing latency testing). References: https://mako.dev/benchmark?benchmark_key=5143375149793280, https://mako.dev/run?run_key=5276035817930752, https://mako.dev/run?run_key=6012349710336000
Eventing moves Channel and Subscription resources to the messaging.knative.dev API group, redefining them as routing constructs rather than simple event forwarders. https://github.com/knative/eventing/pull/1742
Open‑Source Project Recommendations
microk8s – a lightweight, single‑package Kubernetes distribution ideal for offline development, IoT, and edge devices. Includes serverless (knative), service mesh (Istio), monitoring (Prometheus, Grafana), and ML (Kubeflow). https://github.com/ubuntu/microk8s
qlkube – provides a GraphQL API for Kubernetes by generating a GraphQL schema from the OpenAPI/Swagger spec. https://github.com/qlkube/qlkube
kube‑fzf – combines kubectl with the fuzzy finder fzf to simplify common operations such as searching pods, tailing containers, describing resources, exec’ing into pods, and port‑forwarding. https://github.com/thecasualcoder/kube-fzf
Reading Recommendations
"The State of State in Cloud Native Applications" – explores state handling in cloud‑native apps and future directions. https://thenewstack.io/the-state-of-state-in-cloud-native-applications/
"How Kubernetes Could Orchestrate Machine Learning Pipelines" – discusses using Kubeflow pipelines for ML training and MLOps on Kubernetes. https://thenewstack.io/how-kubernetes-could-orchestrate-machine-learning-pipelines/
"Kubernetes Web UIs in 2019" – surveys Kubernetes UI projects and outlines expectations for future web interfaces. https://srcco.de/posts/kubernetes-web-uis-in-2019.html
"Deep Dive into Helm 3" (Chinese) – detailed analysis of Helm 3’s alpha releases and upcoming features. https://yq.aliyun.com/articles/703943
"Knative Eventing Choice Introduction" (Chinese) – introduces the Choice feature for event filtering in Knative Eventing. https://yq.aliyun.com/articles/716661
"Service Mesh Development Trend (Part 2)" (Chinese) – examines Istio’s Mixer v2, Envoy WebAssembly support, Google Traffic Director, and SMI developments. https://skyao.io/talk/201908-servicemesh-development-trend2/
Signed-in readers can open the original source through BestHub's protected redirect.
This article has been distilled and summarized from source material, then republished for learning and reference. If you believe it infringes your rights, please contactand we will review it promptly.
Alibaba Cloud Native
We publish cloud-native tech news, curate in-depth content, host regular events and live streams, and share Alibaba product and user case studies. Join us to explore and share the cloud-native insights you need.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.
