Package io.micrometer.core.instrument.binder.kafka

Class KafkaStreamsMetrics

java.lang.Object
io.micrometer.core.instrument.binder.kafka.KafkaStreamsMetrics
All Implemented Interfaces:
MeterBinder, AutoCloseable
@Incubating(since="1.4.0") public class KafkaStreamsMetrics extends Object
Kafka Streams metrics binder. This should be closed on application shutdown to clean up resources.

It is based on the Kafka client's metrics() method returning a Metric map.

Meter names have the following convention: kafka.(metric_group).(metric_name)

Since:
1.4.0
See Also:

  • Constructor Details

    • KafkaStreamsMetrics

      public KafkaStreamsMetrics(org.apache.kafka.streams.KafkaStreams kafkaStreams, Iterable<Tag> tags)
      KafkaStreams metrics binder
      Parameters:
      kafkaStreams - instance to be instrumented
      tags - additional tags

    • KafkaStreamsMetrics

      public KafkaStreamsMetrics(org.apache.kafka.streams.KafkaStreams kafkaStreams)
      KafkaStreams metrics binder
      Parameters:
      kafkaStreams - instance to be instrumented

    • KafkaStreamsMetrics

      public KafkaStreamsMetrics(org.apache.kafka.streams.KafkaStreams kafkaStreams, Iterable<Tag> tags, ScheduledExecutorService scheduler)
      KafkaStreams metrics binder. The lifecycle of the custom scheduler passed is the responsibility of the caller. It will not be shut down when this instance is closed. A scheduler can be shared among multiple instances of KafkaStreamsMetrics to reduce resource usage by reducing the number of threads if there will be many instances.
      Parameters:
      kafkaStreams - instance to be instrumented
      tags - additional tags
      scheduler - customer scheduler to run the task that checks and binds metrics
      Since:
      1.14.0

    • KafkaStreamsMetrics

      public KafkaStreamsMetrics(org.apache.kafka.streams.KafkaStreams kafkaStreams, Iterable<Tag> tags, ScheduledExecutorService scheduler, Duration refreshInterval)
      KafkaStreams metrics binder. The lifecycle of the custom scheduler passed is the responsibility of the caller. It will not be shut down when this instance is closed. A scheduler can be shared among multiple instances of KafkaStreamsMetrics to reduce resource usage by reducing the number of threads if there will be many instances.

      The refresh interval governs how frequently Micrometer should call the Kafka Client's Metrics API to discover new metrics to register and discard old ones since the Kafka Client can add/remove/recreate metrics on-the-fly. Please notice that this is not for fetching values for already registered metrics but for updating the list of registered metrics when the Kafka Client adds/removes/recreates them. It is the responsibility of the caller to choose the right value since this process can be expensive and metrics can appear and disappear without being published if the interval is not chosen appropriately.

      Parameters:
      kafkaStreams - instance to be instrumented
      tags - additional tags
      scheduler - customer scheduler to run the task that checks and binds metrics
      refreshInterval - interval of discovering new/removed/recreated metrics by the Kafka Client
      Since:
      1.16.0

  • Method Details