66 lines
1.5 KiB
Markdown
66 lines
1.5 KiB
Markdown
|
# 新疆移动 哪吒Session Records图表数据显示异常
|
|||
|
|
|
|||
|
|
| ID | Creation Date | Assignee | Status |
|
|||
|
|
|----|----------------|----------|--------|
|
|||
|
|
| OMPUB-786 | 2023-01-31T17:14:36.000+0800 | 方顺健 | 已解决 |
|
|||
|
|
|
|||
|
|
|
|||
|
|
---
|
|||
|
|
|
|||
|
|
环境：XJ-移动 哪吒
|
|||
|
|
版本： 22.02版本
|
|||
|
|
问题描述：查看Data Flow页面Session Records图表，发现选取不同时间范围，同一时间点的数值就会不同；如下图所示以2023.01.24 23:30时间点为例，选取最近7天，最近30天，和输入时间范围的数值均不相同，且差异较大；**fangshunjian** commented on *2023-01-31T18:13:05.546+0800*:
|
|||
|
|
|
|||
|
|
NZ根据查询的不同时间范围自动计算 step ，规则如下：
|
|||
|
|
* 0 < interval < 24h, step = 15s
|
|||
|
|
* 24h <= interval < 7days, step = 5min
|
|||
|
|
* 7days <= interval < 30days ,step = 10min
|
|||
|
|
* 30days <= interval , step = 30min
|
|||
|
|
|
|||
|
|
prometheus sum函数计算步骤：
|
|||
|
|
* 根据step 抽样计算点
|
|||
|
|
* 相同时间计算点数据求和
|
|||
|
|
|
|||
|
|
结论：
|
|||
|
|
* 不同时间范围 或 相同时间范围起始时间不一样 sum统计结果不一致，属正常现象
|
|||
|
|
* 可以考虑使用 sum_over_time() 函数统计一段时间内求和
|
|||
|
|
|
|||
|
|
|
|||
|
|
|
|||
|
|
[https://segmentfault.com/a/1190000017553625]
|
|||
|
|
|
|||
|
|
https://prometheus.io/docs/prometheus/latest/querying/operators/
|
|||
|
|
|
|||
|
|
https://zhuanlan.zhihu.com/p/295463393
|
|||
|
|
|
|||
|
|
[https://www.youtube.com/watch?v=67Ulrq6DxwA&t=1826s]
|
|||
|
|
|
|||
|
|
|
|||
|
|
|
|||
|
|
|
|||
|
|
|
|||
|
|
|
|||
|
|
|
|||
|
|
---
|
|||
|
|
|
|||
|
|
|
|||
|
|
|
|||
|
|
## Attachments
|
|||
|
|
|
|||
|
|
**34638/24日当天.png**
|
|||
|
|
|
|||
|
|
---
|
|||
|
|
|
|||
|
|
**34636/输入选择最近7天.png**
|
|||
|
|
|
|||
|
|
---
|
|||
|
|
|
|||
|
|
**34635/最近30天.png**
|
|||
|
|
|
|||
|
|
---
|
|||
|
|
|
|||
|
|
**34637/最近7天.png**
|
|||
|
|
|
|||
|
|
---
|
|||
|
|
|