Cloud Native Enthusiast Weekly: Dapr becomes CNCF Incubation Project

Cloud Native Weekly News:

• Dapr joins CNCF incubator
• Knative 1.0 released
• Canonical begins to provide Ubuntu images optimized for Intel CPUs
• Longhorn brings cloud-native distributed storage to CNCF incubator
• Open source project recommendation
• Article recommendation

Cloud native dynamics

Dapr (distributed application runtime) joins CNCF incubator

A few days ago, CNCF Technical Oversight Committee (TOC) has voted to accept Dapr as a CNCF incubation project.

Dapr is a set of APIs that make it easy for developers to write distributed applications. Dapr runs as a sidecar process next to the application, whether in Kubernetes or any other environment, and provides developers with a set of primitives in the form of publish/subscribe, state management, secret management, event triggers, and service-to-service calls Safe and reliable. With Dapr, developers focus on building business logic rather than infrastructure.

Main ingredients:

• Dapr sidecar-runs alongside the application and contains APIs for developers.
• CLI and SDK-the developer tool experience that makes up the project.
• Components-contrib repository-Developers can extend Dapr to integrate and support various cloud services and open source technologies.

Important milestones:

• 15.1k GitHub Stars
• 1,940 pull requests
• 3,703 issues
• 1.3k contributors
• 14 stable versions v1.4
• 26M Docker pulls

details, see

Knative 1.0 released

A few days ago, Knative 1.0 was released, thanks to the contribution and collaboration of more than 600 developers. The Knative project was released by Google in July 2018. It was developed in close cooperation with VMWare, IBM, Red Hat, and SAP. In the past 3 years, Knative has become the most widely installed serverless layer on Kubernetes.

Knative is composed of many version-controlled components: core components (Serving, Eventing) are all GA (generally available); extended components (service/event functions, net-* plug-ins, channels/agents, sources) will be in Alpha, Beta or GA status.

details, see

Canonical begins to provide Ubuntu images optimized for Intel CPUs

Canonical released the first batch Intel IoT platform to meet the unique needs of intelligent edges across multiple vertical industries.

As the number and scale of IoT deployments continue to expand, maintaining a large number of devices on site has become the main focus of the operations team. Stability and reliability are key pain points that Canonical solves. Canonical integrates the latest Intel kernel patches in different Ubuntu versions, as well as the well-known security and reliability features provided by the containerized Ubuntu Core. In addition, hardware-based security measures integrated into Intel chips help mitigate firmware, code, and data attacks, while dedicated encryption accelerators can accelerate data encryption.

details, see

Longhorn brings cloud-native distributed storage to CNCF incubator

A few days ago, the CNCF Technical Oversight Committee (TOC) has voted to accept Longhorn as a CNCF incubation project.

Longhorn is a distributed block storage system of Kubernetes, designed to run on different types of physical storage devices, infrastructure and architectures. It is built on Kubernetes and is used for workloads running on Kubernetes. Longhorn's control plane is based on the controller design pattern and uses dynamic pod management to provide its data plane communication stack.

Longhorn joined CNCF as a sandbox project in October 2019. Since then, Longhorn has achieved extraordinary growth, from 200 contributors from 30 companies to more than 800 contributors from more than 120 companies. The number of submitters has increased from 14 in 3 companies to more than 70 submitters in more than 13 companies. The number of nodes running Longhorn around the world has increased tenfold, from 2,700 to more than 34,000.

details, see

Open source project recommendation

MangoDB

MongoDB abandoned completely open source and changed its license to SSPL, which made it unusable for other open source and commercial projects. MangoDB is an open source alternative to MongoDB. It converts MongoDB's wire protocol queries into SQL and uses PostgreSQL as the database engine.

kubectl-slice

kubectl-slice is a CLI tool that can cut the Kubernetes configuration list containing multiple object resources into multiple YAML files through rules.

For example, there is such a configuration list:

# example.yaml
apiVersion: v1
kind: Pod
metadata:
  name: nginx-ingress
---
apiVersion: v1
kind: Namespace
metadata:
  name: production

After cutting through kubectl-slice, two separate configuration lists are generated:

$ kubectl-slice --input-file=example.yaml
Wrote pod-nginx-ingress.yaml -- 57 bytes.
Wrote namespace-production.yaml -- 60 bytes.
2 files generated.

Peirates

Peirates is a Kubernetes penetration tool that enables attackers to infiltrate through the Kubernetes cluster by elevating permissions, stealing the Service Account in the cluster, to obtain further code execution permissions, and gain control of the cluster.

PortX

PortX is a cross-platform SSH client, macOS users can use it as an alternative to Xshell.

Article recommendation

KubeSphere 3.2.0 released: Brings GPU scheduling for AI scenarios and a more flexible gateway

What is the most popular server-side technology today? The answer is probably cloud native ! As a cloud-native distributed operating system with Kubernetes as its core, KubeSphere is also a part of this cloud-native boom. KubeSphere continues to uphold the promise of 100% open source, and with the help of the open source community, it quickly goes global.

KubeSphere 3.2.0 brings more exciting features, adding support for "GPU resource scheduling management" and GPU usage monitoring, which further enhances the experience of using cloud native AI scenarios. At the same time, it also enhanced "multi-cluster management, multi-tenant management, observability, DevOps, application store, governance" 16189e44590079 and other features to further improve the interaction design, and comprehensively enhance the user experience.

Efficiently deploy Prometheus federated cluster

Personally, I would rather deploy a set of Thanos in front of the Prometheus cluster instead of manually deploying the Prometheus federated cluster. The author of this article may not have the freedom to deploy other services in the cluster, so he racked his brains to optimize the federation scheme. If this is the case for you, you can take a good look at this article.

Azimo's road to world-class surveillance

Monitoring is divided into three realms:

• Infancy : Monitoring is always one step slower for customers
• Advanced period : The monitoring is consistent with the information obtained by the customer
• Mahayana Period : Monitoring is always one step ahead of customers

Most people stay in the first and second realms. If you want to enter the highest realm, you can take a look at the experience shared by Azimo.

This article is published by the blog one article multi-posting OpenWrite