The delivery problem based on K8s is back! | Exclusive Delivery Cheats (Part 3)

Hello everyone, I am Wang Xiaochui, who focuses on delivery. Students who have read the stories of the Iron Triangle series (my little hammer who focuses on delivery, development brother Tiezi and pre-sales boss Qiangge) must know that the three of us serve in An ISV company that provides big data analysis services often faces various challenges in the application delivery process.

On that day, when we were worried about a delivery problem, someone rushed over excitedly and said they had found a solution. In this way, an exclusive delivery recipe came into our world, and this recipe also claims to make delivery easier when used well. After careful research, we found that the cheats mentioned many tricks to help solve the delivery problem, and one of them impressed us a lot, which is about improving the container base on the basis of the original community version of the container base Kubernetes (hereinafter referred to as K8s). , which can better serve the moves of application delivery. Below, please come with me to see if you are the one chosen for that day!

Chosen

Before learning a move, let's see if you are the chosen one, and it is suitable to use this move for delivery!

Overwhelmed by the heavy workload of adaptation and compatibility?

• Different customers have different environments, and it is necessary to repeatedly configure hardware parameters and operating systems, which increases the cost of environment adaptation. After the POC environment is running, there may still be problems such as system libraries often missing or version incompatibility. And if you encounter a bug in a component, you still need to wait for a long repair support from the community, and you can basically choose an alternative to bypass it. Since the resolution of each problem is time-consuming, it leaves a negative impression of poor product quality.
• A large number of component adaptation work requires manual debugging, which is extremely time-consuming. Use open source K8s and ecological components to build environments and deploy products. Due to the lack of best practices for reference, the configuration flexibility of each component of K8s itself is very high. Many component parameter configurations are not suitable, and people need to debug for themselves. Not only need to install various dependencies and performance testing, but also need to customize each environment.

It is expected that compatibility detection and automatic adaptation of parameters can be realized.

Is the application transformation under the trend of Xinchuang difficult?

• There are more and more scenarios of ARM architecture processor + domestic operating system, and the localization ecosystem is still in the stage of rapid development. It is difficult to find suitable ARM version components. Not only that, the process of finding suitable components and transformation still costs money A lot of time, and there are many requirements for IPv6 network structure or need to support IPv4/IPv6 network dual stack.

It is expected that the ARM version components can be easily found to be compatible with the Xinchuang environment, and to easily solve problems such as network and operating system.

Are you sweating profusely during cluster installation and debugging?

• Due to different environments, a lot of manual configuration modification and debugging work needs to be done when delivering at the customer's site. After the cluster is built, if there is a lack of cluster installation and management tools, developers need to provide a lot of detailed script assistance for life cycle management, which increases the probability of errors during operation and greatly reduces the deployment speed.
• The installation and deployment of the cluster and the expansion and contraction of components are very cumbersome, and the overall operation is not flexible enough. After each installation or expansion is completed, it is necessary to check whether the nodes, network, disk, DNS and other resources are working properly. It is easy to miss the inspection. Many scenarios are faced with the lack of professional tools for monitoring, and one-sided inspection can only be done based on whether the upper-layer application is healthy. Observed.

It is expected to realize life cycle management functions such as automatic deployment, expansion and contraction of minute-level clusters. Provides node environment inspection and cluster health inspection capabilities to achieve agility and ease of use, saving time and effort.

Problem handling and operation and maintenance control are not efficient enough?

• There is a lack of comprehensive cluster installation and operation and maintenance management tools. The monitoring, logging, and alarming capabilities provided by the integration of cloud-native technologies or piecing together to form operation and maintenance tools are often worried that they are not stable because the operation and maintenance services come from many parties. If you build your own operation and maintenance system, the R&D cycle will be prolonged and the delivery cost will increase.
• After building the operation and maintenance tool, the physical resource occupation becomes higher. While ensuring the operation and maintenance capability, it is hoped to control the physical resource occupation to a very small range, reduce the privatization output cost of business products, and ensure the reliability of the operating environment. Products are more competitive.
• Data collection and alerting for cluster monitoring are expensive. After using K8s, if you want to monitor the workload or ensure the best state, you need to collect and understand a large number of indicators at the system level to compile alarms for different scenarios, which is a huge challenge for the application side's research and development, and it takes a lot of time.

It is expected that an automated operation and maintenance management and control platform will solve the above problems.

Introduction

If you happen to encounter the above problems, or have been troubled by the above problems, then you are the person chosen that day, promise me to continue reading to learn about this move.

In recent years, with the development of cloud-native concepts, more and more ISVs choose K8s for application development and deployment planning in the form of cloud-native. The community version of K8s serves as a base with high stability, good compatibility, safety and reliability , has been able to provide good service to customers, and the ISV feels good in actual use and delivery of business. But in increasingly complex customer scenarios, how to make the base better and more stable and reliable?

As mentioned above, although the community version of K8s provides reliable components for everyone to use, can these components provide services that are more suitable for customers' unique scenarios when dealing with complex business situations? When there are so many operation and maintenance capabilities to choose from, how to integrate a comprehensive and efficient operation and maintenance management and control platform to provide to customers, this is the question we are thinking about.

It is our goal to provide customers with differentiated and value-added services that are different from the community version of K8s, and to help customers solve more delivery problems. Fortunately, through continuous exploration, we finally found the direction. Following the concept of customer first, and with technical polishing, we are inspired to create an easy-to-use, stable and reliable container base. Finally, in the cloud-native application delivery platform ADP (hereinafter referred to as ADP), which is built for the application delivery field, we provide ADP-Local, the container base of ADP. Based on native K8s technology, ADP-Local provides necessary system support for the operation, management and maintenance of upper-layer business applications, including k8s, log collection, monitoring alarms, fault diagnosis, etc., and provides a reliable operating environment and operation and maintenance tools for the software. To help customers deal with the difficulties and dilemmas described above easily and effectively.

Winning reason

As we all know, the community version of K8s has been delivered to customer sites by a large number of ISVs as part of the application, and has undergone many verifications. The bases we provide have also experienced the practice of many complex customer scenarios, so you can use them with confidence. Since ADP-Local is a container base created by in-depth research and continuous polishing of many different complex site environments, when you use ADP-Local, you will find that ADP-Local has the following characteristics:

1. Safe and reliable : All core components come from Alibaba Cloud Container Service ACK and are kept synchronously updated. These components have undergone rigorous production verification by hundreds of thousands of commercial users and Alibaba Group's core business scenarios, and their security and reliability have been tested in practice.
2. Diversity and compatibility : The core components support both X86 and ARM hardware architectures. At the same time, the high-performance network plug-in Hybridnet included in the container service ACK distribution makes it possible to diversify the network environment, and ultimately ensures smooth operation in diverse infrastructures. superior.
3. Agile and easy to use: Container Service ACK release is deeply integrated with Alibaba's open-source cluster application packaging and delivery tool Sealer, which implements cluster lifecycle management functions such as automatic cluster deployment, scaling, and upgrading in minutes, and provides node environment inspection and capacity planning. ability.
4. Efficient operation and maintenance : integrated monitoring, alarm, log, backup, diagnosis and other services to help products quickly find and deal with problems, and contribute to the stable operation of products. It also has built-in backup, recovery, upgrade, and disaster recovery solutions, turning various operation and maintenance services that originally required a lot of work into the infrastructure that supports product operation in the delivery process, saving worry and effort.

How to pass the trick

Learn more about ADP - Local

ADP-Local integrates embedded K8s and a number of components that serve the efficient deployment and stable operation of products.

It can serve multiple industries and provide rich product capabilities and platform atomic capabilities, namely embedded K8s. Realize that its products can be more stable, safe, lightweight, elastic, integrated, and tailorable, effectively solving your delivery problems.

After many core business scenarios of Alibaba Group and the complex environment practice and verification of countless users.

• You can use the well-built operation and maintenance management and control platform provided by ADP-Local to detect the compatibility between different environments, and quickly and automatically adapt parameters to reduce your workload.
• At the same time, ADP-Local integrates rich Xinchuang architecture components to help you easily find ARM version components to be compatible with Xinchuang environment, and quickly complete the adaptation and transformation of Xinchuang environment.
• Through ADP-Local, life cycle management such as automatic deployment, expansion and contraction of minute-level clusters can be realized, and reliable node environment inspection and cluster health inspection capabilities are provided.

You can easily find and use these capabilities in ADP-Local, an automated operation and maintenance management and control platform, to help you solve problems in one stop.

You have a preliminary understanding of the capabilities of ADP-Local. At the same time, due to space reasons and more abundant operation and maintenance capabilities, the ADP-Local operation and maintenance moves will be introduced in the next chapter. Click to read the original text to learn more about the cloud native application delivery platform ADP!

Seeing this, I remembered the problems that we delivered to the Iron Triangle to solve through this cheat book, and I couldn't help but fall into memory. Well, today's story will be told here first. The story of the cloud native application delivery platform ADP and us will continue, so stay tuned. If you have any questions, please leave a message below~

(The above stories are purely fictitious, and any similarity is purely coincidental.)

past review

Dismantling of exclusive delivery cheats (first round)

Don't leave home, get delivery - exclusive delivery cheats (second time)