云原生-SKyWalking

Service Apdex（数字）:当前服务的评分
Successful Rate（数字）：请求成功率
Service Load (calls / min) 数字:  分钟请求数
Service Avg Response Times（ms）：平均响应延时，单位ms
Service Apdex（折线图）：一段时间内Apdex评分
Service Response Time Percentile (ms)折线图：服务响应时间百分比
Service Load (calls / min) 折线图:  分钟请求数
Success Rate (%)折线图：分钟请求成功百分比
Message Queue Consuming Count(折线图)：消息队列消耗计数
Message Queue Avg Consuming Latency (ms)折线图：消息队列平均消耗延迟（毫秒）
Service Instances Load (calls / min)：节点请求次数
Slow Service Instance (ms)：每个服务实例(物理机、云主机、pod)的最大延时
Service Instance Success Rate (%)：每个服务实例的请求成功率
Endpoint Load in Current Service (calls / min)：每个端点(URL)的请求次数
Slow Endpoints in Current Service (ms)：当前端点(URL)的最慢响应时间
Success Rate in Current Service (%)：当前服务成功率（%）：

Service Instance Load (calls / min）：当前实例的每分钟请求数。
Service Instance Success Rate (%)：当前实例的请求成功率。
Service Instance Latency (ms)：当前实例的响应延时。
Database Connection Pool：数据库连接池信息
Thread Pool：线程池信息

Endpoints： URL
Load (calls / min)：平均请求次数(默认时间范围半小时)，比如半小时内总请求次数6次，6%30=0.20
Success Rate (%)：平均成功率(默认时间范围半小时)
Latency (ms)：平均延迟时间(默认时间范围半小时)

JVM CPU (%)：jvm占用CPU的百分比。
JVM Memory (MB)：JVM内存占用大小，单位m，包括堆内存，与堆外内存（直接内存）。
JVM GC Time (ms)：JVM垃圾回收时间，包含YGC和OGC。
JVM GC Count：JVM垃圾回收次数，包含YGC和OGC
JVM Thread Count：JVM线程计数统计
JVM Thread State Count：JVM线程状态计
JVM Class Count：JVM类计数

# java -javaagent:/skywalking-agent/skywalking-agent.jar \
	DSW_AGENT_NAMESPACE=xyz \
	DSW_AGENT_NAME=abc-application \
	Dskywalking.collector.backend_service=skywalking.abc.xyz.com:11800 \
	jar abc-xyz-1.0-SNAPSHOT.jar

# 部署tomcat:
# 下载并部署tomcat，
$ wget https://archive.apache.org/dist/tomcat/tomcat-8/v8.5.84/bin/apache-tomcat-8.5.84.tar.gz
$ tar xvf apache-tomcat-8.5.84.tar.gz
$ vim /apps/apache-tomcat-8.5.84/bin/catalina.sh
CATALINA_OPTS="$CATALINA_OPTS -javaagent:/data/skywalking-agent/skywalking-agent.jar"; export CATALINA_OPTS

# 配置skywalking-agent参数：
agent.service_name=${SW_AGENT_NAME:magedu}
agent.namespace=${SW_AGENT_NAMESPACE:jenkins}
collector.backend_service=${SW_AGENT_COLLECTOR_BACKEND_SERVICES:172.31.2.161:11800}

# 上传Jenkins app并启动tomcat：
$ root@skywalking-agent1:/apps# ls /apps/apache-tomcat-8.5.84/webapps/
ROOT  docs  examples  host-manager  jenkins.war  manager
$ root@skywalking-agent1:/apps# /apps/apache-tomcat-8.5.84/bin/catalina.sh  run

pip install "apache-skywalking"

# 设置环境变量
$ export SW_AGENT_NAME='python-app1'
$ export SW_AGENT_NAMESPACE='python-project'
$ export SW_AGENT_COLLECTOR_BACKEND_SERVICES='172.31.2.161:11800'

# 启动项目
$ sw-python -d run python3 manage.py runserver 172.31.4.1:80

# gun启动
sw-python run -p gunicorn your_app:app --workers 2 --worker-class uvicorn.workers.UvicornWorker --bind 0.0.0.0:8088

# uwsgi
sw-python run -p uwsgi --die-on-term --http 0.0.0.0:5000 --http-manage-expect --master --workers 3 --enable-threads --threads 3 --manage-script-name --mount /=main:app

# 下载代理程序
https://skywalking.apache.org/downloads/#GoAgent

# 在项目的根目录中执行命令。此命令将下载 skywalking-go 所需的依赖项
go get github.com/apache/skywalking-go

# 也可以通过命令自动导入
skywalking-go/bin/skywalking-go-agent--darwin-amd64 -inject path/to/your-project


# 将依赖包导入到main文件中
package main

import (
	_ "github.com/apache/skywalking-go"
)

# 编译文件
go build -toolexec="/path/to/go-agent" -a -o test .

root@skywalking-server:~# cat /apps/skywalking/config/oal/core.oal
	service_resp_time #服务的响应时间
	service_sla #服务的http请求成功率SLA,比如99%等。
	service_cpm  #表示每分钟的吞吐量.
	service_apdex : 应用性能指数是0.8是0.x
	service_percentile: 指定最近多少数据范围内的响应时间百分比,即p99, p95, p90, p75, p50在内的数据统计结果
	
	endpoint_relation_cpm #端点的每分钟的吞吐量
	endpoint_relation_resp_time #端点的响应时间
	endpoint_relation_sla  #端点的http请求成功率SLA,比如99%等。
	endpoint_relation_percentile  ##端点的最近多少数据范围内的响应时间百分比,即p99、p95、p90、p75、p50在内的数据统计结果
	endpoint_mq_consume_latency #消息延迟
	
	service_instance_sla #实例的请求成功率SLA,比如99%等
	service_instance_resp_time #实例的响应时间
	service_instance_cpm #实例的每分钟吞吐量

rules:  #定义rule规则
	 service_cpm_rule: #唯一的规则名称,必须以_rule结尾
     # Metrics value need to be long, double or int
     ##metrics-name: service_cpm  #指标名称
     ##op: ">" #操作符,>, >=, <, <=, ==
     ##threshold: 1 #指标阈值
     # The length of time to evaluate the metrics
     ##period: 2 #评估指标的时间长度
     # How many times after the metrics match the condition, will trigger alarm
     ##count: 1 #匹配成功多少次就会触发告警
     # How many times of checks, the alarm keeps silence after alarm triggered, default as same as period.
     ##silence-period: 2 #触发告警后的静默时间
     expression: sum(service_cpm > 10) >= 2 #sum(service_cpm > 1)为计算表达式(每分钟计算一次是否匹配),后面的10为阈值、表示service_cpm的值大于10(等于之前版本的op对比操作符),后面的>=2为2次及以上匹配成功,等于之前版本的count
      period: 10 #The length of time to evaluate    the metrics,评估指标的时间长度,即匹配最近10分钟内的指标数据        silence-period: 1 #后期告发送前静默1分钟
 message: '当前service {name} 分钟访问大于阈值100'

# root@skywalking-oap-698bcb77f9-qj57s:/skywalking# cat config/alarm-settings.yml

rules:
  # Rule unique name, must be ended with `_rule`.
  endpoint_percent_rule:
    # A MQE expression and the root operation of the expression must be a Compare Operation.
    expression: sum((endpoint_sla / 100) < 75) >= 3
    # The length of time to evaluate the metrics
    period: 10
    # How many times of checks, the alarm keeps silence after alarm triggered, default as same as period.
    silence-period: 10
    message: Successful rate of endpoint {name} is lower than 75%
    tags:
      level: WARNING
  service_percent_rule:
    expression: sum((service_sla / 100) < 85) >= 4
    # [Optional] Default, match all services in this metrics
    include-names:
      - service_a
      - service_b
    exclude-names:
      - service_c
    period: 10
    message: Service {name} successful rate is less than 85%
  service_resp_time_percentile_rule:
    expression: sum(service_percentile{_='0,1,2,3,4'} > 1000) >= 3
    period: 10
    silence-period: 5
    message: Percentile response time of service {name} alarm in 3 minutes of last 10 minutes, due to more than one condition of p50 > 1000, p75 > 1000, p90 > 1000, p95 > 1000, p99 > 1000
  meter_service_status_code_rule:
    expression: sum(aggregate_labels(meter_status_code{_='4xx,5xx'},sum) > 10) > 3
    period: 10
    count: 3
    silence-period: 5
    message: The request number of entity {name} 4xx and 5xx status is more than expected.
    hooks:
      - "slack.custom1"
      - "pagerduty.custom1"
  comp_rule:
    expression: (avg(service_sla / 100) > 80) * (avg(service_percentile{_='0'}) > 1000) == 1
    period: 10
    message: Service {name} avg successful rate is less than 80% and P50 of avg response time is over 1000ms in last 10 minutes.
    tags:
      level: CRITICAL
    hooks:
      - "slack.default"
      - "slack.custom1"
      - "pagerduty.custom1"

wechat:
  default:
    is-default: true
    text-template: |-
      {
        "msgtype": "text",
        "text": {
          "content": "Apache SkyWalking Alarm: \n %s."
        }
      }      
    webhooks:
    - https://qyapi.weixin.qq.com/cgi-bin/webhook/send?key=dummy_key

dingtalk:
  default:
    is-default: true
    text-template: |-
      {
        "msgtype": "text",
        "text": {
          "content": "Apache SkyWalking Alarm: \n %s."
        }
      }      
    webhooks:
    - url: https://oapi.dingtalk.com/robot/send?access_token=dummy_token
      secret: dummysecret

feishu:
  default:
    is-default: true
    text-template: |-
      {
        "msg_type": "text",
        "content": {
          "text": "Apache SkyWalking Alarm: \n %s."
        },
        "ats":"feishu_user_id_1,feishu_user_id_2"
      }      
    webhooks:
    - url: https://open.feishu.cn/open-apis/bot/v2/hook/dummy_token
      secret: dummysecret