prometheus java api 读取数据 prometheus查询语句

转载

mob64ca14147fe3 2023-10-12 23:56:15

文章标签 prometheus 时间序列返回结果运算符 文章分类 Java 后端开发

20201028prometheus查询语句学习

https://prometheus.fuckcloudnative.io/di-san-zhang-prometheus/di-4-jie-cha-xun/examples

简单的示例

简单的时间序列选择

返回度量指标prometheus_http_requests_total的所有时间序列样本数据

prometheus_http_requests_total

#返回内容
prometheus_http_requests_total{code="200",handler="/api/v1/label/:name/values",instance="localhost:9090",job="prometheus"}	1
prometheus_http_requests_total{code="200",handler="/api/v1/query",instance="localhost:9090",job="prometheus"}	15
prometheus_http_requests_total{code="200",handler="/graph",instance="localhost:9090",job="prometheus"}	1
prometheus_http_requests_total{code="200",handler="/metrics",instance="localhost:9090",job="prometheus"}	7
prometheus_http_requests_total{code="200",handler="/static/*filepath",instance="localhost:9090",job="prometheus"}	2

返回度量指标为prometheus_http_requests_total，且标签包含job="prometheus"和handler="/metrics"的所有时间序列样本数据：

prometheus_http_requests_total{job="prometheus",handler="/metrics"}

#返回结果
prometheus_http_requests_total{code="200",handler="/metrics",instance="localhost:9090",job="prometheus"}	13

返回度量指标为prometheus_http_requests_total，且标签包含job="prometheus"和handler="/metrics"的最近2分钟内的序列样本数据：

prometheus_http_requests_total{job="prometheus",handler="/metrics"}[2m]

#返回结果 15s收集一次，2min收集8次，没毛病
#注意：这是一个区间向量，不能在graph页面展示
prometheus_http_requests_total{code="200",handler="/metrics",instance="localhost:9090",job="prometheus"}	26 @1603868399.722
27 @1603868414.722
28 @1603868429.722
29 @1603868444.722
30 @1603868459.722
31 @1603868474.722
32 @1603868489.722
33 @1603868504.722

使用正则表达式：匹配标签handler的值是/api开头的所有时间序列

prometheus_http_requests_total{handler=~"/api.*"}

#返回结果
prometheus_http_requests_total{code="200",handler="/api/v1/label/:name/values",instance="localhost:9090",job="prometheus"}	1
prometheus_http_requests_total{code="200",handler="/api/v1/query",instance="localhost:9090",job="prometheus"}	22
prometheus_http_requests_total{code="400",handler="/api/v1/query_range",instance="localhost:9090",job="prometheus"}	1

使用正则表达式：返回指标名称是prometheus_http_requests_total，且http状态码不是200的所有时间序列

prometheus_http_requests_total{code!~"200"}

#返回结果
prometheus_http_requests_total{code="400",handler="/api/v1/query_range",instance="localhost:9090",job="prometheus"}	1

使用函数，操作符等

返回指标prometheus_http_requests_total过去5分钟内的http请求平均增长率（(最后一个值 - 第一个值)/300s）：

rate(prometheus_http_requests_total{handler="/metrics"}[5m])

#返回结果  这个返回结果的单位是 请求个数/每秒
{code="200",handler="/metrics",instance="localhost:9090",job="prometheus"}	0.06666666666666667

返回度量指标prometheus_http_requests_total过去5分钟内的http请求数的平均增长率之和，维度是job （通俗点的理解：将prometheus_http_requests_total该指标过去5分钟的时间序列按job名字分组，每组下面的时间序列算出平均增长速率并求和，得到每个组的平均增长率）

sum(rate(prometheus_http_requests_total[5m])) by (job)

#结果如下 没办法，所有时间序列的标签都是job="prometheus",因为我们只部署了一个prometheus
{job="prometheus"}	0.09473684210526315

返回node节点已使用的内存，单位是MiB，并向下取整

floor((node_memory_MemTotal_bytes - node_memory_MemFree_bytes) /1024 /1024)

#返回结果
{instance="172.17.0.3:9100",job="node-exporter-1"}	1637

初识PromQL

表达式语言数据类型

在Prometheus的表达式语言中，表达式或子表达式包括以下四种类型之一：

瞬时向量（instant vector）：一组时间序列，每个时间序列包含一个样本数据，他们共享相同是时间戳
区间向量（Range vector）：一组时间序列，每个试卷序列包含一段时间范围的样本数据
标量（Scalar）：一个浮点型的数据值
字符串（String）：一个简单的字符串

时间序列过滤器

瞬时向量过滤器（花括号）

通过标签来进行过滤，支持正则表达式，例如返回指标名称是prometheus_http_requests_total，且http状态码不是200的所有时间序列

prometheus_http_requests_total{code!~"200"}

运算符说明：

= 等于
!= 不等于
=~ 正则匹配
!~ 正则不匹配

所有的 PromQL 表达式必须至少包含一个指标名称，或者一个不会匹配到空字符串的标签过滤器。

区间向量过滤器（方括号）

获取某一段时间内的时间序列。例如返回指标prometheus_http_requests_total过去5分钟内的http请求平均增长率

prometheus_http_requests_total[5m]

时间单位说明：

s 秒
m 分
h 小时
d 天
w 周
y 年

时间位移操作（offset）

以当前时间为基准，获取一段时间之前的样本数据。例如 返回指标prometheus_http_requests_total5分钟前的样本数据

prometheus_http_requests_total offset 5m

还可以对区间向量操作，例如 返回指标prometheus_http_requests_total5分钟之前 5分钟内的样本数据（有点绕）

prometheus_http_requests_total[5m] offset 5m

操作符

二元运算符

算术二元运算符

加 +
减 -
乘 *
除 /
取余 %
幂 ^

布尔运算符

== 相等
!= 不等
大于 >
小于 <
大于等于 >=
小于等于 <=

加布尔运算符与不加布尔运算符的区别：

#表达式1
prometheus_http_requests_total{handler="/metrics"} offset 5s > 100
#返回结果1
prometheus_http_requests_total{code="200",handler="/metrics",instance="localhost:9090",job="prometheus"}	305

#表达式2
prometheus_http_requests_total{handler="/metrics"} offset 5s > bool 100
#返回结果2
{code="200",handler="/metrics",instance="localhost:9090",job="prometheus"}	1

集合运算符

and （交集）
or （并集）
unless 排除（差集）

匹配模式

https://prometheus.fuckcloudnative.io/di-san-zhang-prometheus/di-4-jie-cha-xun/operators

两边的样本标签不一致，如何处理等

一对一

一对多

聚合操作

sum 求和
min 求最小值
max 求最大值
avg 求平均值
stddev 标准差
stdvar 标准差异
count 计数
count_values 对value进行计数
bottomk 样本最小的k个元素
topk 样本最大的k个元素
quantile 分布统计

这些操作符被用于聚合所有标签维度，或者通过without或者by子语句来保留不通的维度

without用于从计算结果中移除列举的标签，而保留其他标签。by正好相反，结果向量中只保留列出的标签，其他标签则移除。通过without和by可以按照样本的问题对数据进行聚合。

二元运算符的优先级

1. ^
2. * / %
3. + -
4.== != > < >= <=
5.and unless
6.or

PromQL内置函数

absent 如果指标有样本数据，就返回no data，没有样本数据，就返回1
ceil 向上取整
count 计算个数
floor向下取整
histogram_quantile 计算分位数
increase 获取增长量
irate 瞬时增长率。通过最后两个样本计算，rate函数是通过第一个和最后一个样本计算
label_replace 为了能够让客户端的图标更具有可读性，可以通过label_replace函数为时间序列添加额外的标签

#具体参数
label_replace(v instant-vector, dst_label string, replacement string, src_label string, regex string)
#案例
label_replace(up, "host", "$1", "instance",  "(.*):.*")
#该函数会依次对up指标里的每一条时间序列进行处理，通过正则表达式"(.*):.*"匹配所有标签，将instance字符串替换为host
#输出结果
up{host="localhost",instance="localhost:9090",job="prometheus"}   1

predict_linear 预测时间序列 v 在 t 秒后的值
rate 平均增长率
sort 排序
sum 求和
_over_time()

avg_over_time(range-vector) : 区间向量内每个度量指标的平均值。
min_over_time(range-vector) : 区间向量内每个度量指标的最小值。
max_over_time(range-vector) : 区间向量内每个度量指标的最大值。
sum_over_time(range-vector) : 区间向量内每个度量指标的求和。
count_over_time(range-vector) : 区间向量内每个度量指标的样本数据个数。
quantile_over_time(scalar, range-vector) : 区间向量内每个度量指标的样本数据值分位数，φ-quantile (0 ≤ φ ≤ 1)。
stddev_over_time(range-vector) : 区间向量内每个度量指标的总体标准差。
stdvar_over_time(range-vector) : 区间向量内每个度量指标的总体标准方差。

HTTP API中使用PromQL

Prometheus当前稳定的HTTP API可以通过/api/v1访问

API响应格式

Prometheus API使用了JSON格式的响应内容。调用成功返回2XX状态码。失败状态码有下面几种：

404 Bad Request：当参数错误或缺失时
422 Unprocessable Entity：当表达式无法执行时
503 Service Unavailable：当请求超时或被中断时候

JSON格式：

{
  "status": "success" | "error",
  "data": <data>,

  // Only set if status is "error". The data field may still hold
  // additional data.
  "errorType": "<string>",
  "error": "<string>"
}

瞬时数据查询

GET /api/v1/query

参数：

query=: PromQL表达式
time=: 指定用于计算PromQL的时间戳。可选。默认使用当前系统时间
timeout=: 超时设置。可选参数，默认情况下使用-query，timeout的全局设置

计算表达式up当前的计算结果

prometheus java api 读取数据 prometheus查询语句_时间序列

[root@david ~]$ curl -s  'http://127.0.0.1:9090/api/v1/query?query=up' | jq
{
  "status": "success",
  "data": {
    "resultType": "vector",
    "result": [
      {
        "metric": {
          "__name__": "up",
          "group": "production",
          "instance": "172.17.0.5:8080",
          "job": "example-random"
        },
        "value": [
          1603954696.79,
          "1"
        ]
      },
      {
        "metric": {
          "__name__": "up",
          "group": "production",
          "instance": "172.17.0.5:8081",
          "job": "example-random"
        },
        "value": [
          1603954696.79,
          "1"
        ]
      },
      {
        "metric": {
          "__name__": "up",
          "group": "test",
          "instance": "172.17.0.5:8082",
          "job": "example-random"
        },
        "value": [
          1603954696.79,
          "1"
        ]
      },
      {
        "metric": {
          "__name__": "up",
          "instance": "172.17.0.3:9100",
          "job": "node-exporter-1"
        },
        "value": [
          1603954696.79,
          "1"
        ]
      },
      {
        "metric": {
          "__name__": "up",
          "instance": "localhost:9090",
          "job": "prometheus"
        },
        "value": [
          1603954696.79,
          "1"
        ]
      }
    ]
  }
}

响应数据类型

{
  "resultType": "matrix" | "vector" | "scalar" | "string",
  "result": <value>
}

vector：瞬时向量
matrix：区间向量
scalar：标量
string：字符串

区间数据查询

GET /api/v1/query_range

请求参数：

query=：PromQL表达式
tart=：起始时间戳
end=: 结束时间戳
step=：查询步长
timeout：超时设置。可选参数

演示，查询2分钟内的prometheus请求次数，间隔为30s：

#注意，prometheus用utc时间；时间戳用date +%s获取即可
[root@david ~]$ endtime=$(date +%s --utc);starttime=$(date +%s -d "-2 min" --utc)
[root@david ~]$ curl --location --request GET "http://127.0.0.1:9090/api/v1/query_range?query=prometheus_http_requests_total&start=${starttime}&end=${endtime}&step=30"
{"status":"success","data":{"resultType":"matrix","result":[{"metric":{"__name__":"prometheus_http_requests_total","code":"200","handler":"/api/v1/label/:name/values","instance":"localhost:9090","job":"prometheus"},"values":[[1603958295,"1"],[1603958325,"1"],[1603958355,"1"],[1603958385,"1"],[1603958415,"1"]]},{"metric":{"__name__":"prometheus_http_requests_total","code":"200","handler":"/api/v1/query","instance":"localhost:9090","job":"prometheus"},"values":[[1603958295,"34"],[1603958325,"34"],[1603958355,"34"],[1603958385,"34"],[1603958415,"34"]]},{"metric":{"__name__":"prometheus_http_requests_total","code":"200","handler":"/api/v1/query_range","instance":"localhost:9090","job":"prometheus"},"values":[[1603958295,"3"],[1603958325,"4"],[1603958355,"4"],[1603958385,"4"],[1603958415,"4"]]},{"metric":{"__name__":"prometheus_http_requests_total","code":"200","handler":"/metrics","instance":"localhost:9090","job":"prometheus"},"values":[[1603958295,"1273"],[1603958325,"1275"],[1603958355,"1277"],[1603958385,"1279"],[1603958415,"1281"]]},{"metric":{"__name__":"prometheus_http_requests_total","code":"200","handler":"/static/*filepath","instance":"localhost:9090","job":"prometheus"},"values":[[1603958295,"2"],[1603958325,"2"],[1603958355,"2"],[1603958385,"2"],[1603958415,"2"]]},{"metric":{"__name__":"prometheus_http_requests_total","code":"400","handler":"/api/v1/query","instance":"localhost:9090","job":"prometheus"},"values":[[1603958295,"9"],[1603958325,"9"],[1603958355,"9"],[1603958385,"9"],[1603958415,"9"]]},{"metric":{"__name__":"prometheus_http_requests_total","code":"400","handler":"/api/v1/query_range","instance":"localhost:9090","job":"prometheus"},"values":[[1603958295,"7"],[1603958325,"7"],[1603958355,"7"],[1603958385,"7"],[1603958415,"7"]]},{"metric":{"__name__":"prometheus_http_requests_total","code":"404","handler":"/static/*filepath","instance":"localhost:9090","job":"prometheus"},"values":[[1603958295,"2"],[1603958325,"2"],[1603958355,"2"],[1603958385,"2"],[1603958415,"2"]]}]}}

查询元数据

通过标签选择器查询对应的时间序列

GET /api/v1/series

请求参数：

<series_selector>: 表示标签选择器是series_selector。必须至少提供一个match[]` 参数。
start=<rfc3339 | unix_timestamp> : 起始时间戳。
end=<rfc3339 | unix_timestamp> : 结束时间戳。

查询标签值

GET /api/v1/label/<label_name>/values

[root@david ~]$ curl -s 'http://127.0.0.1:9090/api/v1/label/handler/values' | jq
{
  "status": "success",
  "data": [
    "/",
    "/-/reload",
    "/alerts",
    "/api/v1/label/:name/values",
    "/api/v1/query",
    "/api/v1/query_range",
    "/api/v1/series",
    "/graph",
    "/metrics",
    "/static/*filepath",
    "/targets"
  ]
}

获取所有标签名称

[root@david ~]$ curl -s 'http://127.0.0.1:9090/api/v1/labels' | jq
{
  "status": "success",
  "data": [
    "__name__",
    "address",
    "branch",
    "broadcast",
    "call",
    "cause",
    "code",
    "collector",
    "config",
    "cpu",
    "device",
    "dialer_name",
    "domainname",
    "duplex",
    "endpoint",
    "event",
    "fstype",
    "goversion",
    "group",
    "handler",
    "instance",
    "interval",
    "ip",
    "job",
    "le",
    "listener_name",
    "machine",
    "mode",
    "mountpoint",
    "name",
    "nodename",
    "operstate",
    "path",
    "quantile",
    "queue",
    "reason",
    "release",
    "revision",
    "role",
    "rule_group",
    "scrape_job",
    "service",
    "slice",
    "sysname",
    "type",
    "version"
  ]
}

查询目标target

[root@david ~]$ curl -s 'http://127.0.0.1:9090/api/v1/targets' | jq
{
  "status": "success",
  "data": {
    "activeTargets": [
      {
        "discoveredLabels": {
          "__address__": "172.17.0.5:8082",
          "__metrics_path__": "/metrics",
          "__scheme__": "http",
          "group": "test",
          "job": "example-random"
        },
        "labels": {
          "group": "test",
          "instance": "172.17.0.5:8082",
          "job": "example-random"
        },
        "scrapePool": "example-random",
        "scrapeUrl": "http://172.17.0.5:8082/metrics",
        "globalUrl": "http://172.17.0.5:8082/metrics",
        "lastError": "",
        "lastScrape": "2020-10-29T08:16:43.5058997Z",
        "lastScrapeDuration": 0.0018196,
        "health": "up"
      },
      {
        "discoveredLabels": {
          "__address__": "172.17.0.5:8080",
          "__metrics_path__": "/metrics",
          "__scheme__": "http",
          "group": "production",
          "job": "example-random"
        },
        "labels": {
          "group": "production",
          "instance": "172.17.0.5:8080",
          "job": "example-random"
        },
        "scrapePool": "example-random",
        "scrapeUrl": "http://172.17.0.5:8080/metrics",
        "globalUrl": "http://172.17.0.5:8080/metrics",
        "lastError": "",
        "lastScrape": "2020-10-29T08:16:40.5601044Z",
        "lastScrapeDuration": 0.0102266,
        "health": "up"
      },
      {
        "discoveredLabels": {
          "__address__": "172.17.0.5:8081",
          "__metrics_path__": "/metrics",
          "__scheme__": "http",
          "group": "production",
          "job": "example-random"
        },
        "labels": {
          "group": "production",
          "instance": "172.17.0.5:8081",
          "job": "example-random"
        },
        "scrapePool": "example-random",
        "scrapeUrl": "http://172.17.0.5:8081/metrics",
        "globalUrl": "http://172.17.0.5:8081/metrics",
        "lastError": "",
        "lastScrape": "2020-10-29T08:16:44.3559787Z",
        "lastScrapeDuration": 0.0029326,
        "health": "up"
      },
      {
        "discoveredLabels": {
          "__address__": "172.17.0.3:9100",
          "__metrics_path__": "/metrics",
          "__scheme__": "http",
          "job": "node-exporter-1"
        },
        "labels": {
          "instance": "172.17.0.3:9100",
          "job": "node-exporter-1"
        },
        "scrapePool": "node-exporter-1",
        "scrapeUrl": "http://172.17.0.3:9100/metrics",
        "globalUrl": "http://172.17.0.3:9100/metrics",
        "lastError": "",
        "lastScrape": "2020-10-29T08:16:32.9358176Z",
        "lastScrapeDuration": 0.0147601,
        "health": "up"
      },
      {
        "discoveredLabels": {
          "__address__": "localhost:9090",
          "__metrics_path__": "/metrics",
          "__scheme__": "http",
          "job": "prometheus"
        },
        "labels": {
          "instance": "localhost:9090",
          "job": "prometheus"
        },
        "scrapePool": "prometheus",
        "scrapeUrl": "http://localhost:9090/metrics",
        "globalUrl": "http://40c6141831f7:9090/metrics",
        "lastError": "",
        "lastScrape": "2020-10-29T08:16:44.7169385Z",
        "lastScrapeDuration": 0.003988,
        "health": "up"
      }
    ],
    "droppedTargets": []
  }
}

查村prometheus服务器运行信息

[root@david ~]$ curl -s 'http://127.0.0.1:9090/api/v1/status/runtimeinfo' | jq
{
  "status": "success",
  "data": {
    "startTime": "2020-10-29T02:39:47.5187166Z",
    "CWD": "/prometheus",
    "reloadConfigSuccess": true,
    "lastConfigTime": "2020-10-29T02:39:47Z",
    "corruptionCount": 0,
    "goroutineCount": 41,
    "GOMAXPROCS": 2,
    "GOGC": "",
    "GODEBUG": "",
    "storageRetention": "15d"
  }
}

返回TSDB相关的一些信息

[root@david ~]$ curl -s 'http://127.0.0.1:9090/api/v1/status/tsdb' | jq
{
  "status": "success",
  "data": {
    "headStats": {
      "numSeries": 1640,
      "chunkCount": 9886,
      "minTime": 1603951200000,
      "maxTime": 1603959549668
    },
    "seriesCountByMetricName": [
      {
        "name": "prometheus_http_request_duration_seconds_bucket",
        "value": 90
      },
      {
        "name": "prometheus_http_response_size_bytes_bucket",
        "value": 81
        ......

规则

[root@david ~]$ curl -s 'http://127.0.0.1:9090/api/v1/rules' | jq
{
  "status": "success",
  "data": {
    "groups": [
      {
        "name": "example",
        "file": "/etc/prometheus/prometheus.rules.yml",
        "rules": [
          {
            "name": "job_service:rpc_durations_seconds_count:avg_rate5m",
            "query": "avg by(job, service) (rate(rpc_durations_seconds_count[5m]))",
            "health": "ok",
            "evaluationTime": 0.0005387,
            "lastEvaluation": "2020-10-29T08:22:09.7348335Z",
            "type": "recording"
          }
        ],
        "interval": 15,
        "evaluationTime": 0.0005477,
        "lastEvaluation": "2020-10-29T08:22:09.7348282Z"
      }
    ]
  }
}

警报

[root@david ~]$ curl -s 'http://127.0.0.1:9090/api/v1/alerts' | jq
{
  "status": "success",
  "data": {
    "alerts": []
  }
}

本文由博客群发一文多发等运营工具平台 OpenWrite 发布

本文章为转载内容，我们尊重原作者对文章享有的著作权。如有内容错误或侵权问题，欢迎原作者联系我们进行内容更正或删除文章。

上一篇：用python实现生产计划排产自动化 python自动化流程

下一篇：pymysql 不关闭游标 plsql 游标使用

提问和评论都可以，用心的回复会被更多人看到评论

发布评论

相关文章

官方博客	全部文章	热门标签	班级博客
了解我们	网站地图	意见反馈

鸿蒙开发者社区	51CTO学堂
51CTO	软考资讯