平均内存使用率查询-Prometheus

如何编写一个查询来输出过去24小时内实例的平均内存使用率？ 以下查询显示当前内存使用情况

100 * (1 - ((node_memory_MemFree + node_memory_Cached + node_memory_Buffers) / node_memory_MemTotal))

对于CPU，我可以使用irate

100 * (1 - avg(irate(node_cpu[24h])) BY (instance))

如何使用irate和avg作为内存

过去24小时内实例的平均内存使用率

您可以使用随时间变化的平均值：

100 * (1 - ((avg_over_time(node_memory_MemFree[24h]) + avg_over_time(node_memory_Cached[24h]) + avg_over_time(node_memory_Buffers[24h])) / avg_over_time(node_memory_MemTotal[24h])))

对于CPU，我可以使用irate

100 * (1 - avg(irate(node_cpu[24h])) BY (instance))

irate

只查看最后两个示例，该查询与您拥有的模式数成反比，并且将保持不变（在我的内核中始终为0.1）。你想要

100 - (avg by (instance) (rate(node_cpu{job="node",mode="idle"}[5m])) * 100)

请注意，这是一个5分钟的移动平均线，您可以将

[5m]

更改为您要查找的任何时间段，例如

[24小时]

更新命令

CPU使用率：

(1 - avg(irate(node_cpu_seconds_total{mode="idle"}[10m])) by (instance)) * 100

100 * (1 - ((avg_over_time(node_memory_MemFree_bytes[10m]) + avg_over_time(node_memory_Cached_bytes[10m]) + avg_over_time(node_memory_Buffers_bytes[10m])) / avg_over_time(node_memory_MemTotal_bytes[10m])))

内存使用情况：

(1 - avg(irate(node_cpu_seconds_total{mode="idle"}[10m])) by (instance)) * 100

100 * (1 - ((avg_over_time(node_memory_MemFree_bytes[10m]) + avg_over_time(node_memory_Cached_bytes[10m]) + avg_over_time(node_memory_Buffers_bytes[10m])) / avg_over_time(node_memory_MemTotal_bytes[10m])))

---

此外，我们还可以查询与帐户

Windows的CPU使用率

#Average
100 - (avg by (instance) (irate(windows_cpu_time_total{job=~"dev",mode="idle"}[2m])) * 100)
#Minimum
100 - (max by (instance) (irate(windows_cpu_time_total{job=~"dev",mode="idle"}[2m])) * 100)
#Maximum
100 - (min by (instance) (irate(windows_cpu_time_total{job=~"dev",mode="idle"}[2m])) * 100)

#Average
100.0 - 100 * avg_over_time(windows_os_physical_memory_free_bytes{job=~"prod"}[30d]) / avg_over_time(windows_cs_physical_memory_bytes{job=~"prod"}[30d])
#Maximum
100.0 - 100 * min_over_time(windows_os_physical_memory_free_bytes{job=~"prod"}[30d]) / min_over_time(windows_cs_physical_memory_bytes{job=~"prod"}[30d])
#Minimum
100.0 - 100 * max_over_time(windows_os_physical_memory_free_bytes{job=~"prod"}[30d]) / max_over_time(windows_cs_physical_memory_bytes{job=~"prod"}[30d])

Linux的CPU使用率

#Average
100 - (avg by (instance) (irate(node_cpu_seconds_total{job="$job",mode="idle"}[5m])) * 100)
#Minimum
100 - (max by (instance) (irate(node_cpu_seconds_total{job="$job",mode="idle"}[5m])) * 100)
#Maximum
100 - (min by (instance) (irate(node_cpu_seconds_total{job="$job",mode="idle"}[5m])) * 100)

#Avarege:
100 * (1 - ((avg_over_time(node_memory_MemFree_bytes{job=~"prod"}[30d]) + avg_over_time(node_memory_Cached_bytes{job=~"prod"}[30d]) + avg_over_time(node_memory_Buffers_bytes{job=~"prod"}[30d])) / avg_over_time(node_memory_MemTotal_bytes{job=~"prod"}[30d])))

#Maximum:
100 * (1 - ((min_over_time(node_memory_MemFree_bytes{job=~"prod"}[30d]) + min_over_time(node_memory_Cached_bytes{job=~"prod"}[30d]) + min_over_time(node_memory_Buffers_bytes{job=~"prod"}[30d])) / min_over_time(node_memory_MemTotal_bytes{job=~"prod"}[30d])))

#Minimum
100 * (1 - ((max_over_time(node_memory_MemFree_bytes{job=~"prod"}[30d]) + max_over_time(node_memory_Cached_bytes{job=~"prod"}[30d]) + max_over_time(node_memory_Buffers_bytes{job=~"prod"}[30d])) / max_over_time(node_memory_MemTotal_bytes{job=~"prod"}[30d])))

Windows的内存使用情况

#Average
100 - (avg by (instance) (irate(windows_cpu_time_total{job=~"dev",mode="idle"}[2m])) * 100)
#Minimum
100 - (max by (instance) (irate(windows_cpu_time_total{job=~"dev",mode="idle"}[2m])) * 100)
#Maximum
100 - (min by (instance) (irate(windows_cpu_time_total{job=~"dev",mode="idle"}[2m])) * 100)

#Average
100.0 - 100 * avg_over_time(windows_os_physical_memory_free_bytes{job=~"prod"}[30d]) / avg_over_time(windows_cs_physical_memory_bytes{job=~"prod"}[30d])
#Maximum
100.0 - 100 * min_over_time(windows_os_physical_memory_free_bytes{job=~"prod"}[30d]) / min_over_time(windows_cs_physical_memory_bytes{job=~"prod"}[30d])
#Minimum
100.0 - 100 * max_over_time(windows_os_physical_memory_free_bytes{job=~"prod"}[30d]) / max_over_time(windows_cs_physical_memory_bytes{job=~"prod"}[30d])

Linux的内存使用情况

#Average
100 - (avg by (instance) (irate(node_cpu_seconds_total{job="$job",mode="idle"}[5m])) * 100)
#Minimum
100 - (max by (instance) (irate(node_cpu_seconds_total{job="$job",mode="idle"}[5m])) * 100)
#Maximum
100 - (min by (instance) (irate(node_cpu_seconds_total{job="$job",mode="idle"}[5m])) * 100)

#Avarege:
100 * (1 - ((avg_over_time(node_memory_MemFree_bytes{job=~"prod"}[30d]) + avg_over_time(node_memory_Cached_bytes{job=~"prod"}[30d]) + avg_over_time(node_memory_Buffers_bytes{job=~"prod"}[30d])) / avg_over_time(node_memory_MemTotal_bytes{job=~"prod"}[30d])))

#Maximum:
100 * (1 - ((min_over_time(node_memory_MemFree_bytes{job=~"prod"}[30d]) + min_over_time(node_memory_Cached_bytes{job=~"prod"}[30d]) + min_over_time(node_memory_Buffers_bytes{job=~"prod"}[30d])) / min_over_time(node_memory_MemTotal_bytes{job=~"prod"}[30d])))

#Minimum
100 * (1 - ((max_over_time(node_memory_MemFree_bytes{job=~"prod"}[30d]) + max_over_time(node_memory_Cached_bytes{job=~"prod"}[30d]) + max_over_time(node_memory_Buffers_bytes{job=~"prod"}[30d])) / max_over_time(node_memory_MemTotal_bytes{job=~"prod"}[30d])))

注意：-{job=~“prod”}，是帐户名。您可以使用您的帐户名