你能用promtool去检查你的rules是否正确。

  1. # For a single test file.
  2. ./promtool test rules test.yml
  4. # If you have multiple test files, say test1.yml,test2.yml,test2.yml
  5. ./promtool test rules test1.yml test2.yml test3.yml

测试文件的格式

  1. # This is a list of rule files to consider for testing. Globs are supported.
  2. rule_files:
  3.   [ - <file_name> ]
  5. [ evaluation_interval: <duration> | default = 1m ]
  7. # The order in which group names are listed below will be the order of evaluation of
  8. # rule groups (at a given evaluation time). The order is guaranteed only for the groups mentioned below.
  9. # All the groups need not be mentioned below.
  10. group_eval_order:
  11.   [ - <group_name> ]
  13. # All the tests are listed here.
  14. tests:
  15.   [ - <test_group> ]

<test_group>

  1. # Series data
  2. interval: <duration>
  3. input_series:
  4.   [ - <series> ]
  6. # Name of the test group
  7. [ name: <string> ]
  9. # Unit tests for the above data.
  11. # Unit tests for alerting rules. We consider the alerting rules from the input file.
  12. alert_rule_test:
  13.   [ - <alert_test_case> ]
  15. # Unit tests for PromQL expressions.
  16. promql_expr_test:
  17.   [ - <promql_test_case> ]
  19. # External labels accessible to the alert template.
  20. external_labels:
  21.   [ <labelname>: <string> ... ]

<series> 

  1. # This follows the usual series notation '<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. # This uses expanding notation.
  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>
普罗米修斯允许您为不同的警报规则使用相同的警报名称。因此，在这个单元测试中，您必须在单个下列出alertname的所有触发警报的联合。

  1. # The time elapsed from time=0s when the alerts have to be checked.
  2. eval_time: <duration>
  4. # Name of the alert to be tested.
  5. alertname: <string>
  7. # List of expected alerts which are firing under the given alertname at
  8. # given evaluation time. If you want to test if an alerting rule should
  9. # not be firing, then you can mention the above fields and leave 'exp_alerts' empty.
  10. exp_alerts:
  11.   [ - <alert> ]

<alert>

  1. # Expression to evaluate
  2. expr: <string>
  4. # The time elapsed from time=0s when the expression has to be evaluated.
  5. eval_time: <duration>
  7. # Expected samples at the given evaluation time.
  8. exp_samples:
  9.   [ - <sample> ]

<sample>

  1. # Labels of the sample in usual series notation '<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. # The expected value of the PromQL expression.
  8. value: <number>

示例

这是一个单元测试的示例文件。test.yml是遵循上述语法的测试文件，其中alerts.yml包含报警规则。请和alerts.yml在同一个目录下，运行./promtool test rules test.yml。
test.yml 

  1. # This is the main input for unit testing.
  2. # Only this file is passed as command line argument.
  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

alerts.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."