规则的单元测试

您可以使用promtool测试您的规则。

  1. # 单独测试一个文件
  2. ./promtool test rules test.yml
  4. # 如果你有多个测试文件, test1.yml,test2.yml,test2.yml
  5. ./promtool test rules test1.yml test2.yml test3.yml

测试文件格式 

  1. # 要进行测试的规则文件的列表。支持globs。
  2. rule_files:
  3.   [ - <file_name> ]
  5. # 可选的， default = 1m
  6. evaluation_interval: <duration>
  8. # 下面列出的组名的顺序僵尸规则组的执行顺序(在给定的执行时间内)。仅针对以下提及的组保证顺序。不需要体积所有组
  9. group_eval_order:
  10.   [ - <group_name> ]
  12. # 所有测试均在此处列出
  13. tests:
  14.   [ - <test_group> ]

<test_group> 

  1. # 数据序列
  2. interval: <duration>
  3. input_series:
  4.   [ - <series> ]
  6. # 对上述数据进行单元测试
  8. # 告警规则的单元测试。我们考虑输入文件中的告警规则
  9. alert_rule_test:
  10.   [ - <alert_test_case> ]
  12. # PromQL 表达式的单元测试
  13. promql_expr_test:
  14.   [ - <promql_test_case> ]
  16. # 告警模版可访问的外部扩展标签
  17. external_labels:
  18.   [ <labelname>: <string> ... ]

<series> 

  1. # 遵循常规的序列符号 '<metric name>{<label name>=<label value>, ...}'
  2. # Examples:
  3. #      series_name{label1="value1", label2="value2"}
  4. #      go_goroutines{job="prometheus", instance="localhost:9090"}
  5. series: <string>
  7. # 使用扩展符号
  8. # Expanding notation:
  9. #     'a+bxc' becomes 'a a+b a+(2*b) a+(3*b) … a+(c*b)'
  10. #     'a-bxc' becomes 'a a-b a-(2*b) a-(3*b) … a-(c*b)'
  11. # Examples:
  12. #     1. '-2+4x3' becomes '-2 2 6 10'
  13. #     2. ' 1-2x4' becomes '1 -1 -3 -5 -7'
  14. values: <string>

<alert_test_case>

Prometheus 允许您为不通的告警规则使用相同的告警名称。因此，在此单元测试中，您必须在单个<alert_test_case>下列出警报名称的所有触发告警的并集。

  1. # 检查告警的时间间隔
  2. eval_time: <duration>
  4. # 要测试的告警名称
  5. alertname: <string>
  7. # 给定执行时间以给定警报名称触发的预期告警列表。 如果要测试是否不应该触发警报规则，则可以提及上述字段，并将 "exp_alerts" 留空。
  8. exp_alerts:
  9.  [ - <alert> ]

<alert> 

  1. # 预期告警的扩展标签和注解.
  2. # 注意: 标签包括与告警关联的样本的标签(与您在 `/alerts` 中看到的标签相同，但不包含 `__name__` 和 `alertname` 序列)
  3. exp_labels:
  4.   [ <labelname>: <string> ]
  5. exp_annotations:
  6.   [ <labelname>: <string> ]
  1. # 执行的表达式
  2. expr: <string>
  4. # 执行表达式的时间间隔
  5. eval_time: <duration>
  7. # 给定评估时间的预期样本
  8. exp_samples:
  9.   [ - <sample> ]

<sample> 

  1. # 数据样本的标签采用常规系列标记 '<metric name>{<label name>=<label value>, ...}'
  2. # Examples:
  3. #      series_name{label1="value1", label2="value2"}
  4. #      go_goroutines{job="prometheus", instance="localhost:9090"}
  5. labels: <string>
  7. # PromQL 表达式的期望值.
  8. value: <number>

示例

这是通过测试的单元测试的示例输入文件。test.yml是遵循上述语法的测试文件，alert.yml包含警报规则。

将alerts.yml放在同一目录中，运行./promtool test rules test.yml.

test.yml

  1. # 这是单元测试的入口
  2. # 仅将此文件作为命令行参数传递
  4. rule_files:
  5.     - alerts.yml
  7. evaluation_interval: 1m
  9. tests:
  10.     # Test 1.
  11.     - interval: 1m
  12.       # Series data.
  13.       input_series:
  14.           - series: 'up{job="prometheus", instance="localhost:9090"}'
  15.             values: '0 0 0 0 0 0 0 0 0 0 0 0 0 0 0'
  16.           - series: 'up{job="node_exporter", instance="localhost:9100"}'
  17.             values: '1+0x6 0 0 0 0 0 0 0 0' # 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0
  18.           - series: 'go_goroutines{job="prometheus", instance="localhost:9090"}'
  19.             values: '10+10x2 30+20x5' # 10 20 30 30 50 70 90 110 130
  20.           - series: 'go_goroutines{job="node_exporter", instance="localhost:9100"}'
  21.             values: '10+10x7 10+30x4' # 10 20 30 40 50 60 70 80 10 40 70 100 130
  23.       # Unit test for alerting rules.
  24.       alert_rule_test:
  25.           # Unit test 1.
  26.           - eval_time: 10m
  27.             alertname: InstanceDown
  28.             exp_alerts:
  29.                 # Alert 1.
  30.                 - exp_labels:
  31.                       severity: page
  32.                       instance: localhost:9090
  33.                       job: prometheus
  34.                   exp_annotations:
  35.                       summary: "Instance localhost:9090 down"
  36.                       description: "localhost:9090 of job prometheus has been down for more than 5 minutes."
  37.       # Unit tests for promql expressions.
  38.       promql_expr_test:
  39.           # Unit test 1.
  40.           - expr: go_goroutines > 5
  41.             eval_time: 4m
  42.             exp_samples:
  43.                 # Sample 1.
  44.                 - labels: 'go_goroutines{job="prometheus",instance="localhost:9090"}'
  45.                   value: 50
  46.                 # Sample 2.
  47.                 - labels: 'go_goroutines{job="node_exporter",instance="localhost:9100"}'
  48.                   value: 50

alert.yml

  1. # This is the rules file.
  3. groups:
  4. - name: example
  5.   rules:
  7.   - alert: InstanceDown
  8.     expr: up == 0
  9.     for: 5m
  10.     labels:
  11.         severity: page
  12.     annotations:
  13.         summary: "Instance {{ $labels.instance }} down"
  14.         description: "{{ $labels.instance }} of job {{ $labels.job }} has been down for more than 5 minutes."
  16.   - alert: AnotherInstanceDown
  17.     expr: up == 0
  18.     for: 10m
  19.     labels:
  20.         severity: page
  21.     annotations:
  22.         summary: "Instance {{ $labels.instance }} down"
  23.         description: "{{ $labels.instance }} of job {{ $labels.job }} has been down for more than 5 minutes."