如何对一个指标的标签值进行算术运算（除法）

Question

我想使用标签值来计算一个文件的吞吐量

我在推送网关上收到如下消息 avg_treatment_duration{"treatment_name":"A","step":"E1","length":"1000","fileName"="file1"} 0.3s avg_treatment_duration{"treatment_name":"A","step":"E2","length":"1000","fileName"="file1"} 0.6s avg_treatment_duration{"treatment_name":"B","step":"E3","length":"1000","fileName"="file1"} 0.8s avg_treatment_duration{"treatment_name":"B","step":"E4","length":"1000","fileName"="file1"} 0.6s

avg_treatment_duration{"treatment_name":"A","step":"E1","length":"3000","fileName"="file2"} 0.7s avg_treatment_duration{"treatment_name":"A","step":"E2","length":"3000","fileName"="file2"} 0.6s avg_treatment_duration{"treatment_name":"B","step":"E3","length":"3000","fileName"="file2"} 0.3s avg_treatment_duration{"treatment_name":"B","step":"E4","length":"3000","fileName"="file2"} 0.6s

这些值对应于治疗的持续时间。

操作如下

1. 每个文件的持续时间总和
2. ratio：长度/每个文件的持续时间总和（但我怎样才能得到标签值长度） 3 步骤 2 的结果的平均值

目前，我假设我收到值的比率，所以查询是 avg(sum by(fileName) (avg_treatment_duration{}))

是否可以获取每个文件一个指标的标签值并对其进行算术运算？

谢谢！