大促监控工作
监控范围
- 应用服务器(CPU、内存、网络、IO、GC、JVM,容器云还要关注是否重启)
- 数据库服务器(CPU、内存、QPS、QPM、主从延迟,以及以上指标异常时段的其他指标分析)
- 数据库慢SQL统计
- 系统关键业务单据分时段统计
- 高峰时段API调用统计(以新监控平台为准)
- 高峰时段流量组成(以Dynatrace系统为准)
- 每天发现的问题及原因分析,并根据问题及时调整、更新大促方案,作用于下次大促
应用&数据库(服务器)监控
容器云节点监控:
http://mgr-prod.bard.midea.com:55558/d/-Da4MjF7k/jie-dian-zi-yuan-kan-ban?orgId=1&var-origin_prometheus=annto-prd%E9%9B%86%E7%BE%A4&var-Node=All&var-NameSpace=app-c-csp-admin-1345&var-PodIP=All&var-Pod=All&var-key=quota&from=now-12h&to=now
数据库截图链接:
redis监控:
数据库慢SQL监控
多系统MySQL及慢SQL统一监控(支持导出):
单系统MySQL及慢SQL统计监控(支持跳转):
慢SQL综合查询:
流量监控
入口监控:
https://grafana.midea.com/d/uHH-KZw4z/an-de-ru-kou-jian-kong-by-ye-wu?orgId=1&from=now-12h&to=now
专线监控:
https://grafana.midea.com/d/CI7sstQVz/an-de-nginx-wu-liu-ru-kou-liu-liang-tong-ji?orgId=1&from=now-24h&to=now
接口汇总监控:
https://grafana.midea.com/d/wKYPsUQVk/an-de-nginx-wu-liu-ru-kou-liu-liang-tong-ji-v2?orgId=1
流量情况:
API调用及并发监控
统一接口调用排行:
https://monitor.midea.com/monitoring/monitor/view/overview?systemId=c17efc05bd024d28ae35ddd0ca3874c9&nodeId=c17efc05bd024d28ae35ddd0ca3874c9&group=AnDe&ip=all
并发数:
送装业务数据统计脚本
# 接入订单总数
select count(1) from csp_accept where create_time >'2024-05-20 20:00:00'
-- 每天高峰时段接单量统计
select DATE_FORMAT(create_time,'%Y-%m-%d') as '大促高峰日期',count(1) as '高峰时段(0:00~1:00)接单数量'
from csp_accept where create_time between '2024-05-17 00:00:00' and now()
and DATE_FORMAT(create_time,'%H%i%s')<='010000'
group by DATE_FORMAT(create_time,'%Y-%m-%d')
order by 1 asc
-- 每天接单量统计
select DATE_FORMAT(create_time,'%Y-%m-%d') as '大促高峰日期',count(1) as '接单数量'
from csp_accept where create_time between '2024-05-17 00:00:00' and now()
group by DATE_FORMAT(create_time,'%Y-%m-%d')
order by 1 asc
# 每分钟接单数
select DATE_FORMAT(s.create_time,'%Y-%m-%d %H:%i'),count(1) from csp_accept s
where s.create_time >='2024-05-20 20:00:00' and s.create_time<='2024-05-31 20:30:00'
group by DATE_FORMAT(s.create_time,'%Y-%m-%d %H:%i')
# 每分钟站点解析数量
select DATE_FORMAT(s.create_time,'%Y-%m-%d %H:%i'),count(1) from csp_work_head s
where s.create_time >='2024-05-20 20:00:00' and s.create_time<='2024-05-31 20:30:00'
group by DATE_FORMAT(s.create_time,'%Y-%m-%d %H:%i')
# 待分配网点
select count(1) from csp_accept a where a.create_time >='2024-05-20 20:00:00' and a.create_time<='2024-05-20 23:00:00' and accept_state = 10;