stackstorm 17.编写stackstorm的sensor
1 传感器Sensor基础
Sensors: 获取事件发送给StackStorm。Python编写的。
Triggers: 识别即将发送给StackStorm的事件。trigger是一个元组类型和可选的参数。Sensores注册trigger。
传感器组成: python脚本文件和传感器的YAML元数据文件。
传感器元数据文件:
组成:
class_name: 类名，
entry_point: 是python脚本文件名称,
trigger_types: 触发器类型，包含: 触发器名称name, payload_schema:
2 编写传感器元数据文件
文件名为: count_sensor.yaml
内容如下:
---
class_name: "CountSensor"
entry_point: "count_sensor.py"
description: "Simple polling sensor that emits count number."
poll_interval: 5
trigger_types:
-
name: "count"
description: "A count trigger."
payload_schema:
type: "object"
properties:
count:
type: "number"
pythonpath:
type: "string"
name:
type: "string"分析:
传感器 payload_schema中定义的属性，就是生成的数据，生成的数据是一个字典，
传递给后续 规则使用
polling sensor必须有一个配置项: poll_interval
否则可能导致sensor运行异常
3 编写传感器python文件
文件名为: count_sensor.py
内容如下:
from datetime import datetime
import os
from st2reactor.sensor.base import PollingSensor
class CountSensor(PollingSensor):
def __init__(self, sensor_service, config,
poll_interval=5):
super(CountSensor, self).__init__(
sensor_service=sensor_service,
config=config,
poll_interval=poll_interval
)
self.count = None
self.logger = None
def setup(self):
self.count = 1
self.logger = self.sensor_service.get_logger(name=self.__class__.__name__)
def poll(self):
payload = {
"count": self.count,
"pythonpath": os.environ.get("PYTHONPATH", None),
"name": "chao"
}
currTime = str(datetime.now())
info = "############## time: {currTime}, count: {count}, payload: {payload}".format(
currTime=currTime,
count=self.count,
payload=payload,
)
self.logger.info(info)
self.sensor_service.dispatch(trigger="default.count", payload=payload)
self.count = self.count + 1
def cleanup(self):
pass
def add_trigger(self, trigger):
# This method is called when trigger is created
pass
def update_trigger(self, trigger):
# This method is called when trigger is updated
pass
def remove_trigger(self, trigger):
# This method is called when trigger is deleted
pass
4 运行传感器
将传感器的Python文件和YAML元数据文件放到自己需要存放的pack的目录
地址如下:
/opt/stackstorm/packs/
这里放入到:
/opt/stackstorm/packs/default/sensors/目录下
5 注册传感器
执行如下命令:
st2ctl reload --register-sensors
6 编写传感器sensor对应的规则rule
6.1 编写rule
文件名: count_sensor_rule.yaml
内容如下:
---
name: "count_sensor_rule"
pack: "default"
description: "count sensor rule description"
enabled: true
trigger:
type: "default.count"
criteria:
trigger.name:
type: "equals"
pattern: "chao"
action:
ref: default.count_action
parameters:
content: "{{ trigger.name }}"将count_sensor_rule.yaml放在
/opt/stackstorm/packs/default/rules/目录下
6.2 创建rule
执行:
st2 rule create /opt/stackstorm/packs/default/rules/count_sensor_rule.yaml
7 创建rule中对应的action
7.1 编写mistral的元数据文件
文件名为: multi_task_mistral_meta.yaml
内容如下:
---
name: 'multi_task_mistral'
description: 'multi task mistral description'
enabled: true
runner_type: 'mistral-v2'
entry_point: 'workflows/multi_task_mistral.yaml'
parameters:
cmd:
required: true
type: string
timeout:
type: integer
default: 60将multi_task_mistral_meta.yaml
放入到/opt/stackstorm/packs/default/actions目录下
7.2 编写mistral的的脚本文件
文件名为: multi_task_mistral.yaml
内容如下:
version: '2.0'
default.multi_task_mistral:
desctiption: "milti task mistral desctiption"
type: direct
input:
- cmd
- timeout
output:
stdout: <% $.stdout %>
tasks:
task1:
action: core.local cmd=<% $.cmd %> timeout=<% $.timeout %>
publish:
result1: '<% task(task1).result.stdout %>'
stdout: <% task(task1).result.stdout %>
stderr: <% task(task1).result.stderr %>
on-success:
- task2
on-error:
- task4
task2:
action: core.local cmd='llll'
publish:
stdout: <% task(task2).result.stdout %>
stderr: <% task(task2).result.stderr %>
on-success:
- task3
on-error:
- task4
task3:
action: core.local cmd='echo task3'
publish:
stdout: <% task(task3).result.stdout %>
stderr: <% task(task3).result.stderr %>
on-error:
- task4
task4:
action: core.local cmd='echo task4'
publish:
stdout: <% task(task4).result.stdout %>
stderr: <% task(task4).result.stderr %>将multi_task_mistral.yaml
放在/opt/stackstorm/packs/default/actions/workflows目录下
7.3 编写mistral中的action的元数据文件
文件名为: count_sensor_action_meta.yaml
内容如下:
name: "count_action"
description: "count action description"
runner_type: "python-script"
enabled: true
entry_point: "count_sensor_action.py"
parameters:
content:
type: "string"
description: "the content parameter"
required: true
position: 0将count_sensor_action_meta.yaml
放在/opt/stackstorm/packs/default/actions目录下
7.4 编写mistral中的action的脚本文件
文件名为: count_sensor_action.py
内容如下:
from datetime import datetime
from st2common.runners.base_action import Action
class CountAction(Action):
# append content into file
def run(self, content):
fileName = "/home/count.txt"
try:
currTime = str(datetime.now())
# note, a+ means append instead of w+
with open(fileName, "a+") as fr:
realContent = str(content) + " time: " + currTime + "\n"
fr.write(realContent)
info = "content: {content}, time: {time}".format(
content=content,
time=currTime
)
print info
return (True, content)
except Exception as ex:
info = "write content : {content}, exception: {exp}, message: {mes}".format(
content=content,
exp=ex.__class__.__name__,
mes=ex
)
print info
return (False, content)
将count_sensor_action.py放在
/opt/stackstorm/packs/default/actions目录下
7.5 注册mistral和action
7.5.1 先注册action
即执行如下命令:
st2 action create /opt/stackstorm/packs/default/actions/count_sensor_action_meta.yaml
7.5.2 再注册mistral
st2 action create /opt/stackstorm/packs/default/actions/multi_task_mistral_meta.yaml
8 总结
1) stackstorm完整的数据处理过程是:
sensors->triggers->rules->mistrals/actions
2) 编写一个复杂sensor的调试过程如下:
先跑通action/mistral(可以单独执行action或mistral)
然后跑通rule
然后跑通sensor，通过添加日志方式查看信息
最后打通sensor和rule
参考:
[1] https://docs.stackstorm.com/sensors.html
[2] https://docs.stackstorm.com/2.8/reference/packs.html