Sky Walking
==========
SkyWalking: Large-Scale Distributed Systems Tracing Infrastructure, 是一个对JAVA分布式应用程序集群的业务运行情况进行追踪、告警和分析的系统。
[![Build Status](https://travis-ci.org/wu-sheng/sky-walking.svg?branch=master)](https://travis-ci.org/wu-sheng/sky-walking)
![license](https://img.shields.io/aur/license/yaourt.svg)
# 简介 / abstract
* 核心理论为[Google Dapper论文:Dapper, a Large-Scale Distributed Systems Tracing Infrastructure](http://research.google.com/pubs/pub36356.html),英语有困难的同学可参考[国内翻译](http://duanple.blog.163.com/blog/static/70971767201329113141336/)
* 本分析系统能通过动态字节码技术,对现有的JAVA应用或J2EE应用进行监控和数据收集,并针对应用进场进行准实时告警。此外提供大量的调用性能分析功能,解决目前的监控系统主要监控进程、端口而非应用实际性能的问题。
* 支持国内常用的dubbo、dubbox、motan、httpclient等常见RPC框架,支持应用异常的邮件告警
* skywalking-sdk层面提供的埋点API,同步阻塞访问时间小于100μs
* 通过[byte-buddy](https://github.com/raphw/byte-buddy),插件将通过动态字节码机制,避免代码侵入性,完成监控。动态代码模式埋点,同步阻塞访问时间应在200-300μs
* 通过[DataCarrier](https://github.com/wu-sheng/DataCarrier),提供高效的并行处理能力
* 提供一定的日志数据分析和展现能力,减少或者避免使用团队的二次开发
* SkyWalking is an open source Large-Scale Distributed Systems Tracing Infrastructure, also been known as APM(Application Performance Management) tool. SkyWalking provides a solution to help monitor and analysis a Large-Scale Distributed Systems.
* SkyWalking supports popular rpc frameworks, such as [dubbo](https://github.com/alibaba/dubbo), [dubbox](https://github.com/dangdangdotcom/dubbox), [motan](https://github.com/weibocom/motan) etc., supports email-alert when application occurs unexpected exception。
* SkyWalking's basic API, execution time of blocking saving span must less than 100μs.
* By using [byte-buddy](https://github.com/raphw/byte-buddy) (Thanks to [raphw](https://github.com/raphw)), plugins use dynamic byte code generation to avoid invasive codes. plugins API, execution time of blocking saving span must between 200μs and 300μs, including execution time of dynamic byte code.
* By using [DataCarrier](https://github.com/wu-sheng/DataCarrier), provide parallel high-throughput processing capability.
* Provide trace log analysis and presentation capabilities, Reduce or avoid add-on functions development.
# Supported components
* web container
* Tomcat7
* Tomcat8
* database
* mysql
* oracle
* h2
* easily extend to support sybase, sqlserver, jtds, db2, informix
* rpc framework
* dubbo (tested 2.5.3)
* dubbox (tested 2.8.4-dubbo protocol, 2.8.4-rest protocol)
* httpClient 4.x
* motan (tested 0.1.2)
* cache
* jedix 2.x
* java local method
* config package name or class name (Recommend to use in dev or test enviroment. Will cost too many in product enviroment.)
# 主要贡献者 / Contributors
* 吴晟 [wusheng](https://github.com/wu-sheng) wu.sheng@foxmail.com
* 张鑫 [zhangxin](https://github.com/ascrutae)
___
In October 2016, Sky Walking won `OneAPM Open Source Achievement Award`. The award appreciates sky walking for its "*contribution to popularization of APM technology*".
[OneAPM](http://www.oneapm.com/) will provide financial support for the project in 2016-2017.
Thanks all users of sky walking project.
___
# 交流
* Mail to:wu.sheng@foxmail.com
* QQ群:392443393,请注明“Sky Walking交流”
* 谁在使用Sky Walking?[点击进入](https://github.com/wu-sheng/sky-walking/issues/34)。同时请各位使用者反馈下,都在哪些项目中使用。
* if you are using SkyWalking,[Report to us](https://github.com/wu-sheng/sky-walking/issues/34) please.
# 整体架构图
![2.0 架构图](http://wu-sheng.github.io/sky-walking/sample-code/screenshoot/2.0-2016/SkyWalkingArch.jpg)
# 典型页面展现 / Typical UI show
* 支持浏览器:Firefox/Chrome
## 实时调用链路
* 实时链路追踪展现
![追踪连路图1](http://wu-sheng.github.io/sky-walking/sample-code/screenshoot/1.0b/callChain.png)
* 实时链路追踪详细信息查看
![追踪连路图2](http://wu-sheng.github.io/sky-walking/sample-code/screenshoot/1.0b/callChainDetail.png)
* 实时链路追踪日志查看
![追踪连路图3](http://wu-sheng.github.io/sky-walking/sample-code/screenshoot/1.0b/callChainLog.png)
* 实时链路异常告警邮件
![告警邮件](http://wu-sheng.github.io/sky-walking/sample-code/screenshoot/1.0b/alarmMail.png)
* 添加应用
![添加应用](http://wu-sheng.github.io/sky-walking/sample-code/screenshoot/1.0b/addApp.png)
* 应用列表展现
![应用列表展现](http://wu-sheng.github.io/sky-walking/sample-code/screenshoot/1.0b/appList.png)
## 分析汇总
* 分析结果查询,根据viewpoint模糊匹配查询。查询方式为viewpoint: + 关键字
![应用列表展现](http://wu-sheng.github.io/sky-walking/sample-code/screenshoot/1.0b/searchAnalysisResult.png)
* 分析结果展现,通过分析结果查询页面点击进入
![应用列表展现](http://wu-sheng.github.io/sky-walking/sample-code/screenshoot/1.0b/analysisResult.png)
* 分析典型调用链展现
![应用列表展现](http://wu-sheng.github.io/sky-walking/sample-code/screenshoot/1.0b/typicalAnalysisResult.png)
# v2.0-2016 Quick Start
## 部署第三方软件 / Required of third party softwares
- JDK 1.7
- zookeeper 3.4.6
- mysql
- redis-3.0.5
## 插件支持的JDK / Supported jdk version
- 1.7以上版本 / support 1.7+
## 编译与部署 / Build and deploy
- 服务端发布版本[下载](https://github.com/wu-sheng/sky-walking/releases) (.tar.gz)
- Download Server release version. [Download](https://github.com/wu-sheng/sky-walking/releases) (.tar.gz)
- [配置应用服务器](CONFIG_DOC.md)
- [config servers](CONFIG_DOC.md)
- 如果您想自己编译最新版本,可参考《[代码编译说明](BUILD_DOC.md)》
- [Code compilation instructions](BUILD_DOC.md)
## 下载官方的agent / Download offical agent
[ ![Download](https://api.bintray.com/packages/wu-sheng/skywalking/com.a.eye.skywalking-agent/images/download.svg) ](https://bintray.com/wu-sheng/skywalking/com.a.eye.skywalking-agent/_latestVersion)
- 通过github-release[下载](https://github.com/wu-sheng/sky-walking/releases)
- 通过[skywalking bintray官网](https://bintray.com/wu-sheng/skywalking/)托管,可使用公网仓库[https://jcenter.bintray.com/](https://jcenter.bintray.com/)下载。
- use public repository [https://jcenter.bintray.com/](https://jcenter.bintray.com/) to download sdk
## 下载并设置授权文件 / Download auth file
- 注册并登陆过skywalking-webui,创建应用。(一个用户代表一个逻辑集群,一个应用代表一个服务集群。如前后端应用应该设置两个应用,但归属一个用户)
- Sign up and login in skywalking-webui. Create application as needed.
- 下载授权文件,放在在agent jar所在目录中
- download auth file(*.auth), and copy the auth file to directory of the agent jar
## 在运行时环境中设置环境变量 / set environment variables
```
export SKYWALKING_RUN=true
```
- 设置完成后,SkyWalking将随应用启动运行
- After set env, SkyWalking will be working, when application startup.
## 使用-javaagent 并启动应用程序 / start application with -javaagent
- 为应用程序添加启动参数
```shell
java -jar ... -javaagent:/..ospath../skywalking-agent-x.x-Final.jar
```
- Tomcat可以使用修改启动文件,配置-javaagent参数
# 在应用程序中显示traceid / How to find tid
- 通过web应用的http调用入口,通过返回的header信息,找到此次调用的traceid。前提:此web应用的url,已经使用skywalking进行监控。
# Development Documents
- [Development Guide of Plugin](https://github.com/wu-sheng/sky-walking/wiki/Development-Guide-of-Plugin)
# QA
- [SkyWalking SDK是否已经工作? Is SkyWalking SDK Running?](QA/IS_RUNNING.md)
- [tid在web-ui上无法查询. tid can't be search on web-ui](QA/TID_CANNOT_BE_SEARCH.md)
- [SkyWalking Server的运行情况. The status of SkyWalking Server](QA/SERVER_RUNNING_STAUTS.md)
- [SkyWalking Analysis部署运行常见问题. The FAQ of SkyWalking Analysis deployment](QA/deploy-sw_analysis-problem.md)
# 源代码说明
* [网络传输协议、缓存文件协议说明. the protocols of transform between agent and collector, collector buffer file](doc/protocol.md)
# Recent RoadMap
* [In the schedule] v2.1-2017
* release a version following OpenTracing.io
* provide bridge mode lib to integrate data of system and tracer
* provide English doc as default. Also provide Chinese version on Gitbook or Wiki
* support log framework(log4j, log4j2, logback) to output traceid in system log.
* [Dev] v2.0-2016
* 移除对于Hadoop,HBase的依赖,降低部署和运维难度
* 使用内嵌式数据库,和分布式集群
* 移除副本发送机制
* 支持数据文件的时间分片,支持集群TB级别日志文件存储
* 支持统计和高效的准实时统计汇总
* 更多插件的支持
* motan rpc framework (provided)
* [Release] v1.0-Final
* 客户端与服务端间使用扩展后的网络协议,前置的测试版本无法兼容,以提高服务端的稳定性和健壮性。(在内部使用过程中,发生因为安全扫描,导致服务端宕机)
* 引入javaagent模式,无需修改任何代码,完成监控。
* 所有插件将基于字节码技术实现,原有的配置文件模式将被移除。
* [Release] v1.0 beta
* 修复大量前置版本的BUG
* 提供基于MR的分析能力,能够对调用链进行分析汇总,并形成调用链统计视图
* [Release] v1.0 alpha2
* 基于v1.0-alpha1的较稳定版本,在内部环境进行数月运行,修复大量影响服务端运行稳定性的问题。
* v1.0-beta,将提供定时分析任务,部分分析页面展现功能。
* 预计:后续版本将在插件模式上有较大改变,计划针对java api的拦截和第三方拦截提供更小侵入性的方法。配置和拦截模式会发生相应变化。具体模式,参考v1.0-beta发布时的说明。
[More...](https://github.com/wu-sheng/sky-walking/releases)