This repo is Not Required as latest code will be maintained in https://github.com/116davinder/kafka-cluster-ansible/tree/master/roles/nri-kafka

Apache Kafka NewRelic Integration

This is a ansible role which will install newrelic infra jmx plugin and add kafka jmx configs to specified hosts. 

Cautions

1. It can generate atleast half million events per 1min if cluster size is 3.

Assumptions

1. Production Ready Kafka Installation (https://github.com/116davinder/kafka-cluster-ansible )
2. NewRelic Infra Agent should be already installed and version should be higher than 1.3.x.

Requirements

• NewRelic Infra Plugin RPMs.
• JMX Enabled on Kafka Brokers.
• Ansible Management Kafka Cluster.

How to use

1. Include ansible-nri-kafka-jmx role to your playbook.
2. Added following vars
   • kafkaJmxPort
   • kafkaClusterName

Following Stats for Kafka JVM

collect:
# Standard JVM Metrics
    - domain: java.lang
      event_type: kafkaMonitoring
      beans:
          - query: type=GarbageCollector,name=*
            attributes:
                - CollectionCount
                - CollectionTime
                - LastGcInfo
          - query: type=Memory
            attributes:
                - HeapMemoryUsage.Committed
                - HeapMemoryUsage.Init
                - HeapMemoryUsage.Max
                - HeapMemoryUsage.Used
                - NonHeapMemoryUsage.Committed
                - NonHeapMemoryUsage.Init
                - NonHeapMemoryUsage.Max
                - NonHeapMemoryUsage.Used
          - query: type=Threading
            attributes:
                - ThreadCount
                - TotalStartedThreadCount
                - PeakThreadCount
          - query: type=ClassLoading
            attributes:
                - LoadedClassCount
          - query: type=Compilation
            attributes:
                - TotalCompilationTime

Following Stats for Kafka

collect:

    - domain: kafka.controller
      event_type: kafkaMonitoring
      beans:
          - query: type=KafkaController,name=OfflinePartitionsCount
          - query: type=KafkaController,name=ActiveControllerCount
          - query: type=ControllerStats,name=LeaderElectionRateAndTimeMs
          - query: type=ControllerStats,name=UncleanLeaderElectionsPerSec

    - domain: kafka.log
      event_type: kafkaMonitoring
      beans:
          - query: type=LogFlushStats,name=LogFlushRateAndTimeMs

    - domain: kafka.network
      event_type: kafkaMonitoring
      beans:
          - query: type=SocketServer,name=NetworkProcessorAvgIdlePercent
          - query: type=RequestChannel,name=RequestQueueSize
          - query: type=RequestMetrics,name=TotalTimeMs,request=*
          - query: type=RequestMetrics,name=RequestQueueTimeMs,request=*
          - query: type=RequestMetrics,name=LocalTimeMs,request=*
          - query: type=RequestMetrics,name=RemoteTimeMs,request=*
          - query: type=RequestMetrics,name=ResponseQueueTimeMs,request=*
          - query: type=RequestMetrics,name=ResponseSendTimeMs,request=*

    - domain: kafka.server
      event_type: kafkaMonitoring
      beans:
          - query: type=*,name=*
          - query: type=ReplicaFetcherManager,name=MaxLag,clientId=Replica
          - query: type=FetcherLagMetrics,name=ConsumerLag,clientId=*,topic=*,partition=*
          - query: type=DelayedOperationPurgatory,name=PurgatorySize,delayedOperation=*

    - domain: kafka.utils
      event_type: kafkaMonitoring
      beans:
          - query: type=*,name=*

Ref:

• https://docs.confluent.io/current/kafka/monitoring.html