I wish the whole world developer friends, New Year career development tiger roar wind, New Year wage growth, New Year love fortune Tiger Jump Longxiang, New Year code accident under the tiger mouth escape, tiger! The tiger! The tiger!
What is missing from the community is not architectural diagrams, but reusable architectural practices. What the technical team lacks is not open source tools, but solution design ideas based on production environment and business reality.
Bytedance’s business development has gone from zero to one, moving towards massive services and large-scale scenarios. In this process, many ingenious technical practices have been accumulated. This article is a collection of technical practices published on the official account of Bytedance technical team in the past year for readers.
Best practices for Bytedance large-scale buried data governance
This article is compiled from the fourth Meetup of Volcano Engine developer community. It mainly introduces the buried content solution and buried link solution of Bytedance’s flow platform, and reveals how the flow platform supports bytedance’s trillions + real-time data processing.
👉 Review details here
The microservice architecture of Bytedance has evolved
This article is based on a discussion by Cheng Guozhu, leader of Bytedance’s infrastructure/Service Framework team, at QCon 2021. It mainly introduces the technical practice and experience of the Service Framework team on Golang Service Framework and Service Mesh during 2018-2021.
👉 Review details here
ANR optimization practice monitoring tools and analysis ideas
Raster is a monitoring tool developed by Byte. The tool mainly monitors messages in the main thread scheduling process and aggregates them according to certain policies to minimize the impact of the monitoring tool on application performance and memory jitter. At the same time, the message execution process of the four components is monitored to facilitate the tracking and recording of the scheduling and time consumption of such messages. In addition, statistics are made on the messages currently being scheduled and messages to be scheduled in the message queue, so that the overall scheduling situation of the main thread can be played back when problems occur. In addition, we migrate the CheckTime mechanism of system service to the application side and apply it as the thread CheckTime mechanism, so as to predict the system load and scheduling situation in the past period from the timeliness of thread scheduling when the system information is insufficient.
👉 Review details here
Evolution of bytedance 100,000-node HDFS cluster multi-room architecture
HDFS, full name of Hadoop Distributed File System, is a module of The Apache Hadoop project. As the cornerstone of big data storage, HDFS provides massive data storage capability with high throughput. In Bytedance, HDFS provides a service scale of 100,000 nodes per cluster and 10 EXabytes of data per cluster. This article is the specific idea of using HDFS for bytedance optimization.
👉 Review details here
Active experiment of scene in bytedance chaos engineering practice
This paper introduces a key stage of Bytedance in chaos engineering practice: scene active experiment. It is hoped that this paper can help you to deepen the understanding of the value of chaos engineering, and provide more ideas for the design of chaos engineering experiment and the implementation of chaos engineering construction.
👉 Review details here
I’ll show you how to build a frameless buried point system
Generally speaking, a complete buried point system consists of the following three parts: application; Data analysis platform; Data platform SDK. Starting from the whole process of buried point system, engineers from Bytedance technical team will analyze the steps of full link construction of frameless buried point system from shallow to deep.
👉 Review details here
How did Bytedance land on the micro front end
Microfront-end is an architecture similar to microservices. It is an architectural style consisting of multiple independently delivered front-end applications, which are broken down into smaller, simpler applications that can be independently developed, tested, and deployed, but are still cohesive as a single product in the view of users. This article is the case analysis of ByteDance’s landing micro front end, and relevant solutions have been opened source.
👉 Review details here
How does Bytedance do full link pressure measurement?
Full-link pressure test is a process that simulates massive user requests and data to perform pressure tests on the entire service chain based on actual production service scenarios and system environments and continuously tunes the service chain. It is used to detect server performance problems on complex service links based on the full-link pressure test. This paper mainly introduces the full-link pressure measurement system of Bytedance server and the full-link pressure measurement practice of various services of Bytedance.
👉 Review details here
Distributed link tracking in bytedance practice
In the process of development, Bytedance has gradually formed a very complex super-scale micro-service system, which puts forward high requirements for the overall observability solution of the back-end. To solve this problem, the infrastructure intelligent operation and maintenance team developed a link tracking system, integrated and unified massive Metrics/Trace/Log data, and realized a new generation of one-stop full-link observation and diagnosis platform to help services solve problems such as monitoring barriers, link sorting, and performance analysis. This paper will introduce the overall function and technical architecture of Bytedance link tracking system, as well as our thinking and summary in the practice process.
👉 Review details here
Thinking and practice of volcano engine A/B test
Bytedance started using A/B testing when it was founded in 2012, and there’s A saying inside the company that everything can be A/B testing. A/B testing at Bytedance has been A very basic facility and culture. This article is compiled from the fourth session of the Volcano Engine Developer Community Meetup with the same name. It mainly introduces why to do A/B testing, the volcano Engine A/B testing system architecture and best practices.
👉 Review details here