Application hosting means that the developer does not provide a server for the application, but hosts the application to a service provider. After the application is developed, it can be hosted on EDAS.

Function describe Multiple language support EDAS supports hosting Java applications, and the K8s environment also supports hosting multi-language applications such as PHP, Node.js, C and Go. Multiple underlying server options EDAS seamlessly integrates ECS cloud servers and container service Kubernetes versions. You can choose ECS clusters, K8s clusters, and hybrid cloud clusters based on actual hosting needs. Multiple publishing methods Supports the use of console and API to complete creation and deployment, and supports multiple deployment forms of WAR packages, JAR packages and images. Supports application upgrades using the console, API, and plug-ins, and supports single-batch release, batch release, and canary release. Supports the use of Yunxiao and Jenkins to implement CI/CD of applications. Application life cycle management EDAS provides you with full life cycle management services for applications from creation to operation, including application release, start, stop, expansion, reduction, and deletion. You can easily operate and maintain thousands of application instances. Application monitoring EDAS integrates the application real-time monitoring service ARMS. You can monitor the health status and key indicators of applications deployed in EDAS, set alarms for monitoring indicators, and promptly discover and handle possible anomalies or failures to ensure the health and availability of applications. Application operation and maintenance EDAS provides a series of operation and maintenance functions for applications, including life cycle management, application access management, elastic scaling, current limiting and degradation, load balancing, change recording, event center, log management, etc.

Different from the traditional monolithic application architecture, the microservice architecture splits the application into multiple core functions. Each function can be used as an independent service and can be built and deployed independently by an independent team. Each service Operation and failure do not affect each other. Microservices make applications easier to scale and faster to develop, accelerating innovation and shortening the release cycle of new features.

EDAS provides a complete microservice solution from development, deployment to governance.

Cloud native means that applications are designed to run optimally on the cloud and take full advantage of the cloud. In the cloud-native era, the PaaS platform will sink to the infrastructure and become part of the cloud. EDAS is a PaaS platform for application hosting and microservice management.

EDAS supports three mainstream microservice frameworks: Apache Dubbo (version 2.6.x and above), Spring Cloud (Edgware and above), and HSF. For more information, see Spring Cloud Overview, Dubbo Overview, and HSF Overview. The HSF framework built into EDAS is an efficient microservice framework developed by Alibaba, incubated from Alibaba’s best practices in many business scenarios; at the same time, Apache Dubbo and Spring Cloud applications can be migrated to the cloud with zero code intrusion, effectively reducing operation and maintenance costs. It supports multiple advanced features such as lossless application online, offline, canary release, and traffic control, helping you easily build microservice applications on the cloud.

Architecture diagram

Functional advantages

This eliminates the need for operation and maintenance personnel to log in to ECS instances one by one, eliminating the need for cluster maintenance. You only need to log in to the EDAS console to quickly deploy applications through various methods such as WAR packages, JAR packages or mirrors. ECS provides full application life cycle management, including release, rollback, application group management, multi-version coexistence, and Integrated monitoring, logging and other capabilities greatly improve the efficiency of ECS cluster management.

Architecture diagram

Functional advantages

Alibaba Cloud Toolkit is a free plug-in for IntelliJ IDEA, Eclipse and Maven platforms. It is designed to help developers use Alibaba Cloud more conveniently in local development platforms (IDEs) to develop and test efficiently. , diagnose and deploy applications.

Ali-Tomcat is a Servlet container modified based on Apache Tomcat. On the premise of supporting the original core functions, Ali-Tomcat will automatically load the Pandora container when starting to realize the class isolation function.

CPU sharing is a method to improve single-machine resource utilization in Docker technology. For example, for a 2-core 8G machine host, if the CPU sharing ratio is selected to be 1:2, then up to four 1-core 2G Docker instances can be allocated on this host; and so on, select 1:4, You can allocate up to eight 1-core 1G Docker instances. Note that in any case, memory is exclusive and cannot be shared.

Dubbo is an open source distributed service framework that provides high-performance and transparent RPC remote service calling solutions.

Cloud server ECS (Elastic Compute Service) is a simple, efficient and elastically scalable computing service. Help you quickly build more stable and secure applications.

Enterprise Distributed Application Service EDAS (Enterprise Distributed Application Service) is a PaaS platform for application hosting and microservice management. It provides full-stack solutions such as application development, deployment, monitoring, operation and maintenance, and supports micro-services such as Dubbo and Spring Cloud. The service operating environment helps you easily migrate all kinds of applications to the cloud.

EDAS Agent is a Daemon program installed on the user's ECS. It is mainly used for communication between the EDAS service cluster and the applications deployed on the corresponding ECS. During the operation process, it is mainly responsible for application management, status reporting, information acquisition and other functions; it is also the main bridge for information communication between the EDAS console and user applications.

EDAS Container is the basic container for running HSF applications on the EDAS platform. EDAS Container includes Ali-Tomcat and Pandora.

K8s Namespace allocates objects within the system to different Namespaces to form different logically grouped projects, groups or user groups, so that different groups can share and use the entire cluster resources while being managed separately.

Pandora is a lightweight isolation container, which is taobao-hsf.sar. It is used to isolate applications, middleware, and dependencies between middleware so that they do not affect each other. Pandora of EDAS integrates various middleware product plug-ins such as plug-ins for service discovery, plug-ins for configuration push, and plug-ins for call chain tracking. These plug-ins can be used to complete all-round operation and maintenance management of EDAS applications such as service monitoring, governance, tracking, and analysis.

Pandora Boot is a lighter way to use Pandora based on Pandora. Based on Pandora and FatJar technology, the Pandora environment can be started directly in the IDE.

Pod is the smallest deployment unit and billing unit in Kubernetes. It can be composed of one or more containers according to the application scenario. When there are multiple containers in a Pod, these containers will share the Pod's computing resources, storage space, IP, and ports. You can also limit the proportion of computing resources used by each container.

SLB (Server Load Balancer) is a load balancing service that distributes traffic to multiple application instances. The external service capabilities of the application system can be expanded through traffic distribution, and the availability of the application system can be improved by eliminating single points of failure.

Private network VPC (Virtual Private Cloud) is a custom private network created based on Alibaba Cloud. Different private networks are logically isolated at Layer 2. You can create and manage cloud product instances, such as ECS, load balancing, RDS, etc., within the private network you create.

When performing life cycle operations such as application deployment, startup, expansion/reduction, etc. on EDAS, the application change record abstracts the business logic executed during the entire change process into a process (Process) and displays it in a visual way.

A software development practice in which developers regularly merge code changes into a central repository and then run automated builds and tests.

Using JVM bytecode enhancement technology, necessary time-consuming and call sequence records are added to all method calls in the selected method, so that the specific execution sequence during the execution can be viewed.

Call chain query is used to check the call link status in the system, especially slow services and error services.

The EDAS Eagle Eye monitoring system can analyze every system call, message sending and database access of the distributed system to accurately discover system bottlenecks and hidden dangers.

SchedulerX is a distributed task scheduling product developed by Alibaba’s middleware team. Users rely on SchedulerX-Client in their applications and create scheduled tasks in the SchedulerX console. After configuring the corresponding parameters, they can receive the periodic schedule of scheduled tasks by starting the application. The SchedulerX-Server cluster provides high availability and high stability guarantees for scheduling triggering, and can implement distributed scheduling for user client machine clusters.

The paid account refers to the account used to purchase EDAS products. A paid account can be bound to (up to 5) main accounts, and the paid account is also a main account.

The tenant dimension displays the running status of all services of all applications in the current tenant in the past 24 hours, including service call volume, call time, and call error times.

Service query is used to query the services provided or consumed by the target machine.

It can monitor the QPS, response time and error rate of the service in all aspects according to the service call status of the application.

Contrary to service current limiting, each application will call many external services. Configuring downgrade rules for these services can accurately shield inferior services, ensure that the application itself can run stably, and prevent inferior service dependencies from affecting the application's own service capabilities. EDAS configures degradation rules from the response time dimension to help you properly shield low-quality dependencies when dealing with traffic peaks.

EDAS can configure current limiting rules for many services provided by each application to achieve flow control of services and ensure stable operation of services. Current limiting rules can be configured from two dimensions: QPS and threads, helping you to ensure that the system can run smoothly with maximum support capacity when dealing with traffic peaks.

Intuitively understand the mutual calling relationships and related performance data between different services in the form of topology diagrams.

High-speed Service Framework HSF (High-speed Service Framework) is a distributed service framework for enterprise-level Internet architecture. Based on a high-performance network communication framework, it provides services such as service publishing and registration, service invocation, service routing, and service authentication. A series of functional features such as rights, service current limiting, service degradation and service call link tracking.

Health check performs regular checks and reports on containers and applications, and then reports the results to the console, thereby helping you understand the running status of the entire application in the cluster environment and troubleshoot and locate problems.

EDAS collects data from the instance (ECS) where the application is running, and conducts detailed monitoring of the machine's basic indicators such as CPU, memory, load, network and disk. All monitoring is performed on an application-by-application basis for data statistics and processing.

A cluster refers to the combination of cloud resources required to run an application.

Microservice space is used to achieve service isolation. You can use microservice space to isolate different running environments, such as development, testing, and production environments, to avoid affecting service invocation and configuration push in different environments.

In the EDAS console, you can use machine instructions to perform batch operations on ECS instances with Agent installed. Commands can be executed in batches by cluster, application and instance, solving the trouble of repeated operation and maintenance of multiple instances.

An EDAS lightweight configuration and registration center that can be run locally, providing service discovery and configuration management functions.

A component that collects system monitoring logs to generate monitoring data and call chain information. In a VPC network, machines are naturally isolated from servers, and log collectors can open up the connection channel between servers and local machines. Installing a log collector can collect data from all machines in the entire VPC.

The real-time log is a runtime log of the standard output of the Docker container.

The elastic scaling function can sense the status of each server in the cluster (CPU, RT and Load), and realize cluster expansion and contraction in real time based on the status, while ensuring service quality and improving the availability of the cluster system.

Application Real-Time Monitoring Service ARMS (Application Real-Time Monitoring Service) is an Alibaba Cloud application performance management (APM) monitoring product. With ARMS, enterprises can quickly and easily build application monitoring capabilities with second-level response.

Used to monitor the real-time traffic and historical information of the application. Users can use this information to monitor the health status of the application to quickly discover and locate problems.

Application Configuration Management ACM (Application Configuration Management) is a tool product that centrally manages and pushes application configurations in a distributed architecture environment. You can use ACM in EDAS to centrally manage and push application configurations, and you can also isolate and synchronize configurations between different environments based on namespaces.

Application is the basic unit of EDAS management. An application usually contains multiple instances. EDAS provides a complete application life cycle management mechanism, which can complete the comprehensive management of applications from release to operation, including application creation, deployment, startup, rollback, expansion and contraction, and stop and offline operations.

Application instances refer to ECS instances or container instances used to deploy applications. For example, in an ECS cluster, there is only one application instance in an ECS; in a K8s cluster, a Pod is an application instance.

Group all instances (ECS) subordinate to an application so that different versions of the application can be deployed to instances in different groups. EDAS application instance grouping provides the group management function of instances within the application. Users can implement operation and maintenance methods such as Beta release, AB testing, and grayscale release. It supports users to implement application life cycle management, resource monitoring and alarming from a group perspective, which can quickly improve operation and maintenance efficiency.

The environment in which the application runs, such as the EDAS-Container running environment for HSF applications or the running container for open source applications (such as Apache Tomcat).

EDAS provides detailed troubleshooting and performance analysis for applications, including providing JVM heap/non-heap memory, class loading (ClassLoader), threads, Tomcat connector statistics and method tracking on the single machine where the current application is running.

Cloud Service Bus CSB (Cloud Service Bus) is oriented to private clouds and private domains, helping enterprises to achieve cross-system and cross-protocol service capability interoperability between multiple systems of their own, or with systems of partners and third parties. Create a CSB exclusive instance in EDAS to manage and control the external service opening of applications in the target environment. You can also introduce external services and perform management control; open EDAS applications in VPC for use in your own development environment through the public network To test the joint debugging of EDAS applications in Alibaba Cloud VPC.

EDAS is a way of dividing resources. Resource groups can be bound to ECS, clusters, and SLB. Your main account can authorize sub-accounts based on resource groups.

need

solution

need

solution

need

solution