prometheus聚合运算

修改prometheus rules configmap，增加规则：（修改后需重启prometheus）

groups:

    - name: cpu_mem_usage_active

      interval: 30s

      rules:

      - record: cpu_usage_active

        expr: 100 - (avg by (node_name) (irate(node_cpu_seconds_total{mode="idle"}[2m])) * 100)

      - record: mem_usage_active

        expr: 100*(1-node_memory_MemAvailable_bytes/node_memory_MemTotal_bytes)

    - name: cpu-mem-usage-total-avg

      interval: 5m

      rules:

      - record: cpu_utilization_total_avg

        expr: 100 - (avg(irate(node_cpu_seconds_total{mode="idle"}[2m])) * 100)

      - record: mem_utilization_total_avg

        expr: 100*(1-sum(node_memory_MemAvailable_bytes)/sum(node_memory_MemTotal_bytes))

    - name: cpu-usage-5m

      interval: 5m

      rules:

      - record: cpu_usage_max_avg_1h

        expr: max_over_time(cpu_usage_avg_5m[1h])

      - record: cpu_usage_max_avg_1d

        expr: max_over_time(cpu_usage_avg_5m[1d])

    - name: cpu-usage-1m

      interval: 1m

      rules:

      - record: cpu_usage_avg_5m

        expr: avg_over_time(cpu_usage_active[5m])

    - name: mem-usage-5m

      interval: 5m

      rules:

      - record: mem_usage_max_avg_1h

        expr: max_over_time(mem_usage_avg_5m[1h])

      - record: mem_usage_max_avg_1d

        expr: max_over_time(mem_usage_avg_5m[1d])

    - name: mem-usage-1m

      interval: 1m

      rules:

      - record: mem_usage_avg_5m

        expr: avg_over_time(mem_usage_active[5m])