A Kafka-specific metric measures the internal behavior, availability, correctness, or throughput of Kafka components. It complements host-level and JVM-level metrics by showing whether brokers, producers, consumers, and related applications are functioning as intended. section 11.2.2
Use Kafka-specific metrics to expose operational conditions that generic infrastructure monitoring can miss:
- Track consumer lag so applications do not silently fall behind and violate processing-delay or service-level objectives. section 11.2.3
- Observe Kafka client and broker behavior rather than relying only on CPU, disk, network, heap, or garbage-collection indicators. section 11.2.2
- Assign instrumentation jointly to development and operations teams, especially when cloud platforms already collect infrastructure and JVM signals. section 11.2.3
Detecting a silently falling-behind consumer
A Kafka application may remain technically available while its consumer lag grows. Exposing and tracking a Kafka-specific lag metric reveals the processing delay before it becomes an SLA violation. section 11.2.3
