kafka-alerts.yaml

kind: PrometheusRule
apiVersion: monitoring.coreos.com/v1
metadata:
  prometheus: dx
spec:
  groups:
    - name: kafka-alert
      interval: 10m
      rules:
        - expr: avg by(topic, partition, project,namespace,instance,application) (kafka_consumergroup_lag{topic!="__consumer_offsets"}) > 50000
          alert: '[P3]-Kafka消息堆积高'
          for: 5m
          labels:
            severity: info
          annotations:
            description: 实例为 {{ $labels.application }} 的Kafka，topic：{{ $labels.topic }} 消息堆积高于50000条，实际堆积量为{{ humanize $value }}
            level: P3
            ruleGroupName: kafka-alert
            ruleName: [P3]-Kafka消息堆积高
            type: kafka
        - expr: avg by(topic, partition, project,namespace,instance,application) (kafka_consumergroup_lag{topic!="__consumer_offsets"}) > 70000
          alert: [P2]-Kafka消息堆积高
          for: 5m
          labels:
            severity: warn
          annotations:
            description: 实例为 {{ $labels.application }} 的Kafka，topic：{{ $labels.topic }} 消息堆积高于70000条，实际堆积量为{{ humanize $value }}
            level: P2
            ruleGroupName: kafka-alert
            ruleName: [P2]-Kafka消息堆积高
            type: kafka