StackStorm介绍
StackStorm是一个强大的自动化平台,结合DevOps和ChatOps,提供可扩展、灵活和健壮的工具链用于应用、服务和工作流的自动化能力。
StackStorm核心概念
StackStorm的工作步骤大体如下:
1. StackStorm Sensor感应并触发事件。
2. Rules Engine对事件进行规则匹配,如果匹配产生任务。
3. StackStorm Worker执行任务,一般是调用到外部系统。
4. StackStorm记录审计任务执行的细节。
5.任务执行结果返回给Rules Engine进行进一步处理。
可以看出StackStorm是个以事件驱动的系统,为此抽象出一系列概念来分解事件从产生、触发、规则匹配到执行的整个生命周期事件,具体包含核心概论如下:
Sensor感应器
Sensor是一系列的感应器用于接受或者监测事件,当事件发生的时候,Sensor将会通知Trigger提交事件到StackStorm。
Sensor是Python插件实现,只要实现StackStorm定义的接口,然后配置元数据YAML注册到StackStorm:
--- class_name:"SampleSensor" entry_point:"sample_sensor.py" description:"Sample sensor that emits triggers." trigger_types: - name:"event" description:"An example trigger." payload_schema: type:"object" properties: executed_at: type:"string" format:"date-time" default:"2014-07-30 05:04:24.578325" |
Trigger触发器
Trigger代表事件,一般事件是由外部系统产生,比如监控告警,JIRA问题更新等等,另外也有通用的事件触发器,比如定时器或者WebHook。
在StackStorm系统中,Trigger只是String类型的对象,由Sensor注册,用户可以在Sensor插件自定义新的Trigger。
Action 动作/任务
Action是事件触发后的处理方式,一般是由外部系统执行,包括:
- 重启服务
- 创建云服务
- 发生邮件
- 启动Docker容器
- 制作VM快照
Action可以是通用的执行方式,比如SSH,REST API调用,也能够集成Openstack、Docker/Kubernetes等系统实现。Action Runner是Action的执行环境,StackStorm的内置Action Runner:
Action Runner |
Description |
local-shell-cmd |
This is the local runner. This runner executes a Linux command on the same host where StackStorm components are running. |
local-shell-script |
This is the local runner. Actions are implemented as scripts. They are executed on the same hosts where StackStorm components are running. |
remote-shell-cmd |
This is a remote runner. This runner executes a Linux command on one or more remote hosts provided by the user. |
remote-shell-script |
This is a remote runner. Actions are implemented as scripts. They run on one or more remote hosts provided by the user. |
python-script |
This is a Python runner. Actions are implemented as Python classes with arun method. They run locally on the same machine where StackStorm components are running. |
http-request |
HTTP client which performs HTTP requests for running HTTP actions. |
action-chain |
This runner supports executing simple linear work-flows. |
mistral-v2 |
Those runners are built on top of the Mistral OpenStack project and support executing complex work-flows. |
cloudslang |
This runner is built on top of the CloudSlang project and supports executing complex workflows. |
通过ActionRunner用户可以自定义Action的实现,以下是一个python-script类型的Action用于发送SMS:
--- name:"send_sms" runner_type:"python-script" description:"ThissendsanSMSusingtwilio." enabled:true entry_point:"send_sms.py" parameters: from_number: type:"string" description:"Yourtwilio'from'numberinE.164format.Example+14151234567." required:true position:0 to_number: type:"string" description:"RecipientnumberinE.164format.Example+14151234567." required:true position:1 secret:true body: type:"string" description:"Bodyofthemessage." required:true position:2 default:"Hello{%ifsystem.user%}{{system.user}}{%else%}dude{%endif%}!" |
Workflow 工作流
Workflow是Action集合,Workflow能够定义Action的执行顺序和条件,组合一系列Action完成复杂的任务。Workflow可以认为是广义意义上的Action。
StackStorm支持2种类型的Workflow:
- ActionChain:通过简单的语法定义Action链
--- chain: - name: "c1" ref: "core.local" parameters: cmd: "echoc1" on-success: "c2" on-failure: "c4" - name:"c2" ref: "core.local" parameters: cmd: "echo\"c2:parentexecis{{action_context.parent.execution_id}}.\"" on-success: "c3" on-failure: "c4" - name:"c3" ref: "core.local" parameters: cmd: "echoc3" on-failure: "c4" - name:"c4" ref: "core.local" parameters: cmd: "echofailc4" default: "c1" |
- Mistral :Openstack的工作流组件,可以同Stackstorm集成,支持复杂的工作流配置。
version: '2.0' examples.mistral-join: description: > A sample workflow that demonstrates how to join parallel branches. type: direct tasks: a: action: core.local input: cmd: "echo 'a'" on-success: - b - c - d b: action: core.local input: cmd: "echo 'b'" on-success: - e c: action: core.local input: cmd: "echo 'c'" on-success: - e d: action: core.local input: cmd: "echo 'd'" on-success: - e e: join: all action: core.local input: cmd: "echo 'e'" |
Rule 规则
Rule是映射Trigger到Action(或者Workflow),即当事件触发后,通过Rule定义的标准(Criteria)进行匹配,当匹配成功将执行Action(或者Workflow)。
Rule的定义格式:
--- name: "rule_name" # required pack: "examples" # optional description: "Ruledescription." # optional enabled: true # required trigger: # required type: "trigger_type_ref" criteria: # optional trigger.payload_parameter_name1: type: "regex" pattern : "^value$" trigger.payload_parameter_name2: type: "iequals" pattern : "watchevent" action: # required ref: "action_ref" parameters: # optional foo: "bar" baz: "{{trigger.payload_parameter_1}}" |
Audit 审计
Audit是用来跟踪和记录Action的执行细节,用于查询定位:
{ "status": "succeeded", "start_timestamp": "2014-10-31T02:00:46.679000Z", "parameters": { "cmd": "ifconfig" }, "callback": {}, "result": { ... }, "context": { "user": "stanley" }, "action": "core.local", "id": "5452ed4e0640fd6b59e75908" } |
ChatOps
ChatOps是一种新的DevOps方法,ChatOps是诞生于GitHub的一种基于会话驱动的协作开发方法,过去团队之间的通讯和开发操作是两层皮,导致各种不透明和低效率。ChatOps将开发工具带入开发者聊天室,通过定制的插件和脚本,一个聊天机器人能够执行聊天中输入的各种命令,实现在聊天平台上的团队协作开发自动化,把团队沟通和执行统一整合到一个可视化更高的聊天环境中,“聊着天就把事情办了”。
目前流行的ChatOps聊天机器人主要有Hubot(GitHub的bot,用CoffeeScript和Node.js开发)、Lita(用Ruby开发)和Err(用Python开发)三种,都是开源软件,而且可以整合到开发团队在工作中经常会使用一些聊天工具例如HipChat、Slack、Flowdock和Campfire等。
StackStorm中集成了Hubot作为聊天机器人提供ChatOps,同时提供Action Alias 和Notifications 机制实现更好的体验,如下图所示:
StackStorm ChatOps的流程如下:
Phase 1:Bot Initialiazation
聊天机器人会调用StackStorm API下载Action Alias信息进行初始化。
Phase 2:Alias Execution
初始化完成后,聊天机器人将服务在聊天频道等待命令,一旦接受到命令,将Action Alias转化为Action,并发送给StackStorm进行处理。
Phase 3: ActionExecution
StackStorm处理执行Action,将执行输出发送给ActionRunner,并转发给Notifications子系统。
Phase 4: ActionCompetition
Action执行完成后,将返回Notifications。
Phase 5: NotificationsReply
聊天机器人收到Notifications进行回复。
示例:
参考:https://docs.stackstorm.com/_images/chatops_demo.gif
参考
- https://docs.stackstorm.com/index.html
作者简介
吴龙辉,现任网宿科技云计算架构师,致力于云计算PaaS的研究和实践,《Kubernetes实战》作者,活跃于CloudFoundry,Docker,Kubernetes等开源社区,贡献代码和撰写技术文档。邮箱: [email protected]