InfluxDB-telegraf-Ceph-Flux查询语句

1.查询集群fsid列表

from(bucket: "cts")
  |> range(start: v.timeRangeStart, stop: v.timeRangeStop)
  |> filter(fn: (r) => r["_measurement"] == "ceph_cluster_stats")
  |> filter(fn: (r) => r["_field"] == "value")
  |> group(columns: ["fsid"])
  |> distinct(column: "fsid")

2. 查询集群名称列表

需要在telegraf配置文件的tag中增加cluster_name

import "influxdata/influxdb/v1"
v1.tagValues(
    bucket: v.bucket,
    tag: "cluster_name",
    predicate: (r) => true,
    start: -1d
)

3.查看ceph集群使用率 百分比

from(bucket: "cts")
  |> range(start: v.timeRangeStart, stop: v.timeRangeStop)
  |> filter(fn: (r) => r["_measurement"] == "ceph_cluster_stats")
  |> filter(fn: (r) => r["_field"] == "value")
  |> filter(fn: (r) => r["cluster_name"] == "横扫饥饿做回自己")
  |> filter(fn: (r) => r["type_instance"] == "bytes_used" or r["type_instance"] == "bytes_total")

  |> aggregateWindow(every: v.windowPeriod, fn: mean, createEmpty: true)
    |> pivot(columnKey: ["type_instance"], rowKey: ["_time"], valueColumn: "_value")
    |>filter(fn: (r) => r.bytes_total != 0)
    |> map(fn: (r) => ({ _value: r.bytes_used / float(v: r.bytes_total) }))
  |> yield(name: "mean")

4.Ceph健康状态

HEALTH_ERR = 0,
HEALTH_WARN = 1,
HEALTH_OK = 2

from(bucket: "cts")
  |> range(start: v.timeRangeStart, stop: v.timeRangeStop)
  |> filter(fn: (r) => r["_measurement"] == "ceph_cluster_stats")
  |> filter(fn: (r) => r["_field"] == "value")
  |> filter(fn: (r) => r["cluster_name"] == "横扫饥饿做回自己")
  |> filter(fn: (r) => r["type_instance"] == "health")
  |> yield(name: "mean")

5.Ceph总磁盘数

from(bucket: "cts")
  |> range(start: v.timeRangeStart, stop: v.timeRangeStop)
  |> filter(fn: (r) => r["_measurement"] == "ceph_cluster_stats")
  |> filter(fn: (r) => r["_field"] == "value")
  |> filter(fn: (r) => r["cluster_name"] == "${cluster_name}")
  |> filter(fn: (r) => r["type_instance"] == "num_osd")
  |> yield(name: "mean")

5.Ceph活跃盘数

from(bucket: "cts")
  |> range(start: v.timeRangeStart, stop: v.timeRangeStop)
  |> filter(fn: (r) => r["_measurement"] == "ceph_cluster_stats")
  |> filter(fn: (r) => r["_field"] == "value")
  |> filter(fn: (r) => r["cluster_name"] == "${cluster_name}")
  |> filter(fn: (r) => r["type_instance"] == "num_osd_up")
  |> yield(name: "mean")

5.Ceph在线盘数

from(bucket: "cts")
  |> range(start: v.timeRangeStart, stop: v.timeRangeStop)
  |> filter(fn: (r) => r["_measurement"] == "ceph_cluster_stats")
  |> filter(fn: (r) => r["_field"] == "value")
  |> filter(fn: (r) => r["cluster_name"] == "${cluster_name}")
  |> filter(fn: (r) => r["type_instance"] == "num_osd_in")
  |> yield(name: "mean")

6.异常磁盘数

from(bucket: "cts")
  |> range(start: v.timeRangeStart, stop: v.timeRangeStop)
  |> filter(fn: (r) => r["_measurement"] == "ceph_cluster_stats")
  |> filter(fn: (r) => r["_field"] == "value")
  |> filter(fn: (r) => r["cluster_name"] == "${cluster_name}")
  |> filter(fn: (r) => r["type_instance"] == "num_osd" or r["type_instance"] == "num_osd_up")
  |> pivot(columnKey: ["type_instance"], rowKey: ["_time"], valueColumn: "_value")
  |> map(fn: (r) => ({ _value: r.num_osd - r.num_osd_up }))
  |> yield(name: "mean")

7.磁盘读写延时

from(bucket: "cts")
  |> range(start: v.timeRangeStart, stop: v.timeRangeStop)
  |> filter(fn: (r) => r["type_instance"] == "osd.op_latency")
  |> aggregateWindow(every: v.windowPeriod, fn: mean, createEmpty: false)
  |> derivative(unit: v.windowPeriod)
  |> yield(name: "mean")

8.集群iops

from(bucket: "telegraf_ceph")
  |> range(start: v.timeRangeStart, stop: v.timeRangeStop)
  |> filter(fn: (r) => r["_measurement"] == "ceph_pgmap")
  |> filter(fn: (r) => r["_field"] == "read_op_per_sec" or r["_field"] == "write_op_per_sec")
  |> filter(fn: (r) => r["cluster_name"] == "${cluster_name}")
  |> aggregateWindow(every: v.windowPeriod, fn: mean, createEmpty: false)
  |> yield(name: "mean")

9.集群读写延时

from(bucket: "cts")
  |> range(start: v.timeRangeStart, stop: v.timeRangeStop)
  |> filter(fn: (r) => r["type_instance"] == "osd.op_latency")
  |> aggregateWindow(every: v.windowPeriod, fn: mean, createEmpty: false)
  |> derivative(unit: v.windowPeriod)
  |> yield(name: "mean")

10.重构延时

from(bucket: "telegraf_ceph")
  |> range(start: v.timeRangeStart, stop: v.timeRangeStop)
  |> filter(fn: (r) => r["_measurement"] == "ceph_pool_stats")
  |> filter(fn: (r) => r["_field"] == "recovering_bytes_per_sec" )
  |> filter(fn: (r) => r["cluster_name"] == "${cluster_name}")
|> group(columns: ["cluster_name","_field","_start","_stop"], mode: "by")
 |> sum(column: "_value")

11.重构速率

from(bucket: "telegraf_ceph")
  |> range(start: v.timeRangeStart, stop: v.timeRangeStop)
  |> filter(fn: (r) => r["_measurement"] == "ceph_pool_stats")
  |> filter(fn: (r) => r["_field"] == "recovering_keys_per_sec" )
  |> filter(fn: (r) => r["cluster_name"] == "${cluster_name}")
|> group(columns: ["cluster_name","_field","_start","_stop"], mode: "by")
 |> sum(column: "_value")

12. PG分布

from(bucket: "cts")
  |> range(start: v.timeRangeStart, stop: v.timeRangeStop)
  |> filter(fn: (r) => r["_measurement"] == "ceph_daemon_stats")
  |> filter(fn: (r) => r["type_instance"] == "osd.numpg")
  |> filter(fn: (r) => r["cluster_name"] == "${cluster_name}")
  |> aggregateWindow(every: v.windowPeriod, fn: mean, createEmpty: false)
  |> yield(name: "mean")