Apache Kafka metrics

  • リリースバージョン: Australia
  • 更新日 2026年03月12日
  • 所要時間:5分
  • The following table lists the metrics that are gathered as output from Kafka checks. Entries indicated as Featured metrics are high-visibility metrics that are displayed in the Operator Workspace Metric tab after an alert is generated. These metrics provide the operator with additional information to help them further explore the specified issue.

    表 : 1. Broker metrics
    Metric type Resource (name of specific database, where relevant) Units Metric type description
    Kafka.Broker.RequestMetrics.TotalTimeMs.FetchConsumer.Mean (featured metric) Total time (in ms) to serve the Consumer request.
    Kafka.Broker.RequestMetrics.TotalTimeMs.FetchFollower.Mean (featured metric) Total time (in ms) to serve the Follower request.
    Kafka.Broker.RequestMetrics.TotalTimeMs.Produce.Mean (featured metric) Total time (in ms) to serve the Producer request.
    Kafka.Broker.BrokerTopicMetrics.BytesOutPerSec.OneMinuteRate (featured metric) Aggregate outgoing byte rate.
    Kafka.Broker.BrokerTopicMetrics.BytesInPerSec.OneMinuteRate (featured metric) Aggregate incoming byte rate.
    Kafka.Broker.ReplicaManager.UnderReplicatedPartitions.Value Number of unreplicated partitions.
    Kafka.Broker.ReplicaManager.IsrExpandsPerSec.OneMinuteRate Rate at which the pool of in-sync replicas (ISRs) shrinks.
    Kafka.Broker.ReplicaManager.IsrShrinksPerSec.OneMinuteRate Rate at which the pool of in-sync replicas (ISRs) shrinks.
    Kafka.Broker.KafkaController.ActiveControllerCount.Value (featured metric) Number of active controllers in cluster.
    Kafka.Broker.KafkaController.OfflinePartitionsCount.Value Number of offline partitions.
    Kafka.Broker.ControllerStats.LeaderElectionRateAndTimeMs.OneMinuteRate (featured metric) Leader election rate and latency.
    Kafka.Broker.ControllerStats.UncleanLeaderElectionsPerSec.OneMinuteRate Number of unclean elections per second.
    Kafka.Broker.DelayedOperationPurgatory.PurgatorySize.Fetch.Value Number of requests waiting in fetch purgatory.
    Kafka.Broker.DelayedOperationPurgatory.PurgatorySize.Produce.Value Number of requests waiting in producer purgatory.
    Kafka.Broker.RequestMetrics.RequestsPerSec.Produce.OneMinuteRate (featured metric) Number of producer requests per second.
    Kafka.Broker.RequestMetrics.RequestsPerSec.FetchConsumer.OneMinuteRate (featured metric) Number of consumer requests per second.
    Kafka.Broker.RequestMetrics.RequestsPerSec.FetchFollower.OneMinuteRate (featured metric) Number of follower requests per second.
    表 : 2. Zookeeper metrics
    Metric type Resource (name of specific database, where relevant) Units Metric type description
    Kafka.Zookeeper.outstanding_requests (featured metric) Number of requests queued.
    Kafka.Zookeeper.avg_latency (featured metric) Average amount of time it takes to respond to a client request (in ms).
    Kafka.Zookeeper.num_alive_connections (featured metric) Number of clients connected to ZooKeeper.
    Kafka.Zookeeper.open_file_descriptor_count Number of file descriptors in use.
    Kafka.Zookeeper.pending_syncs Number of pending syncs from followers.
    Kafka.Zookeeper.followers Number of active followers.