Métricas do Dataproc

O Cloud Monitoring mostra o desempenho, o tempo de atividade e a integridade geral de aplicativos com tecnologia de nuvem. O Google Cloud Observability coleta e processa métricas, eventos e metadados de clusters do Dataproc, incluindo métricas HDFS, YARN, job e operação por cluster, para gerar insights por meio de painéis e gráficos (consulte Métricas do Dataproc do Cloud Monitoring).

Coleta de métricas de recursos do Dataproc

O Cloud Monitoring coleta métricas relacionadas aos seguintes recursos do Dataproc:

  • Cluster do Cloud Dataproc
  • Job do Cloud Dataproc
  • Lote do Cloud Dataproc
  • Sessão do Cloud Dataproc

As métricas de recursos do Dataproc são coletadas no seguinte formato: dataproc.googleapis.com/RESOURCE/METRIC, e incluem a coleta de várias métricas do OSS.

Conferir as métricas de recursos do Dataproc

É possível selecionar e visualizar as métricas do recurso Dataproc no Metrics Explorer digitando "dataproc" na caixa Filter by resource or metric name e selecionando um recurso "Cloud Dataproc".

Coleta de métricas personalizadas

Ao criar um cluster do Dataproc, você pode ativar a coleta de métricas de uma ou mais origens de métrica personalizada. Um conjunto padrão de métricas é coletado de cada origem de métrica ativada, a menos que você especifique as métricas a serem coletadas de uma origem de métrica. As métricas especificadas pelo usuário são chamadas de "substituições".

As métricas personalizadas do OSS são coletadas no seguinte formato: custom.googleapis.com/OSS_COMPONENT/METRIC

Exemplos de métricas personalizadas de OSS: 

custom.googleapis.com/spark/driver/DAGScheduler/job/allJobs
custom.googleapis.com/hiveserver2/memory/MaxNonHeapMemory

Ativar a coleta de métrica personalizada

É possível usar a CLI gcloud ou a API Dataproc para ativar a coleta de métricas personalizadas de uma ou mais origens de métricas.

CLI da gcloud

Coleção de métricas personalizadas

Use a flag gcloud dataproc clusters create --metric-sources para ativar a coleta de métricas personalizadas de uma ou mais fontes de métricas. 

gcloud dataproc clusters create cluster-name \
    --metric-sources=METRIC_SOURCE(s) \
    ... other flags

Observações:

Modificar a coleta de métricas

Opcionalmente, adicione a flag --metric-overrides ou --metric-overrides-file para ativar a coleta de uma ou mais das métricas personalizadas de uma ou mais fontes de métricas.

  • Qualquer uma das métricas personalizadas e todas as métricas do Spark podem ser listadas para coleta como uma substituição de métrica. Os valores de métrica de substituição são sensíveis a maiúsculas e minúsculas e precisam ser fornecidos, se apropriado, no formato CamelCase.

    Exemplos:

    • sparkHistoryServer:JVM:Memory:NonHeapMemoryUsage.committed
    • hiveserver2:JVM:Memory:NonHeapMemoryUsage.used
    • yarn:ResourceManager:JvmMetrics:MemHeapMaxM

  • Somente as métricas substituídas especificadas serão coletadas de uma determinada origem de métricas. Por exemplo, se uma ou mais métricas spark:executive forem listadas como substituições de métricas, outras métricas SPARK não serão coletadas. A coleta de métricas personalizadas de outras fontes não é afetada. Por exemplo, se as fontes de métricas SPARK e YARN estiverem ativadas e as substituições forem fornecidas apenas para métricas do Spark, o conjunto padrão de métricas do YARN ativadas será coletado.
  • A origem da substituição de métrica especificada precisa estar ativada. Por exemplo, se uma ou mais métricas spark:driver forem fornecidas como substituições de métricas, a origem da métrica spark precisa estar ativada (--metric-sources=spark).

Lista de métricas de substituição

gcloud dataproc clusters create cluster-name \
    --metric-sources=METRIC_SOURCE(s) \
    --metric-overrides=LIST_OF_METRIC_OVERRIDES \
    ... other flags

Observações:

  • --metric-sources: obrigatório para ativar a coleta de métrica personalizada. Especifique uma ou mais das seguintes fontes de métricas: spark, flink, hdfs, yarn, spark-history-server, hiveserver2, hivemetastore e monitoring-agent-defaults. O nome da origem da métrica não diferencia maiúsculas de minúsculas. Por exemplo, "yarn" ou "YARN" são aceitáveis.
  • --metric-overrides: forneça uma lista de métricas no seguinte formato:

    METRIC_SOURCE:INSTANCE:GROUP:METRIC

    Exemplo:--metric-overrides=sparkHistoryServer:JVM:Memory:NonHeapMemoryUsage.committed

    Essa flag é uma alternativa à flag --metric-overrides-file e não pode ser usada com ela.

Arquivo de métricas de substituição

gcloud dataproc clusters create cluster-name \
    --metric-sources=METRIC-SOURCE(s) \
    --metric-overrides-file=METRIC_OVERRIDES_FILENAME \
    ... other flags

Observações:

  • --metric-sources: obrigatório para ativar a coleta de métrica personalizada. Especifique uma ou mais das seguintes fontes de métricas: spark, flink, hdfs, yarn, spark-history-server, hiveserver2, hivemetastore e monitoring-agent-defaults. O nome da origem da métrica não diferencia maiúsculas de minúsculas. Por exemplo, "yarn" ou "YARN" são aceitáveis.
  • --metric-overrides-file: especifique um arquivo local ou do Cloud Storage (gs://bucket/filename) que contenha uma ou mais métricas no seguinte formato:

    METRIC_SOURCE:INSTANCE:GROUP:METRIC

    Use o formato CamelCase conforme apropriado.

    Exemplos:

    • --metric-overrides-file=gs://my-bucket/my-filename.txt
    • --metric-overrides-file=./local-directory/local-filename.txt

      Essa flag é uma alternativa à flag --metric-overrides e não pode ser usada com ela.

API REST

Use DataprocMetricConfig como parte de uma solicitação clusters.create para ativar a coleta de métricas personalizadas. Observação: monitoring-agent-defaults não está disponível em clusters da versão 2.2 da imagem, a menos que o agente de operações esteja instalado.

Conferir métricas personalizadas

Para selecionar e visualizar as métricas do recurso do Dataproc no Metrics Explorer, selecione o recurso VM Instance e, em seguida, Custom metrics.

Métricas personalizadas

É possível ativar o Dataproc para coletar as métricas personalizadas listadas nas tabelas a seguir.

  • A coluna Métricas ativadas é marcada com "y" se o Dataproc coletar a métrica quando você ativar a origem de métrica associada.

  • Qualquer uma das métricas listadas para uma origem de métrica e todas as métricas do Spark podem ser ativadas para coleta se você substituir a coleta do conjunto padrão de métricas ativadas para a origem de métrica. Consulte Ativar a coleta de métrica personalizada personalizadas.

  • O Dataproc usa o agente de monitoramento para coletar métricas. Ativar qualquer origem de métrica ativa a coleta de métricas do agente. Essas métricas não são cobradas dos usuários. O Dataproc as usa para diagnosticar problemas de coleta de métricas.

Métricas do Hadoop

Métricas do HDFS

Métrica Nome do Metrics Explorer Métricas ativadas
hdfs:NameNode:FSNamesystem:CapacityTotalGB dfs/FSNamesystem/CapacityTotalGB y
hdfs:NameNode:FSNamesystem:CapacityUsedGB dfs/FSNamesystem/CapacityUsedGB y
hdfs:NameNode:FSNamesystem:CapacityRemainingGB dfs/FSNamesystem/CapacityRemainingGB y
hdfs:NameNode:FSNamesystem:FilesTotal dfs/FSNamesystem/FilesTotal y
hdfs:NameNode:FSNamesystem:MissingBlocks dfs/FSNamesystem/MissingBlocks n
hdfs:NameNode:FSNamesystem:ExpiredHeartbeats dfs/FSNamesystem/ExpiredHeartbeats n
hdfs:NameNode:FSNamesystem:TransactionsSinceLastCheckpoint dfs/FSNamesystem/TransactionsSinceLastCheckpoint n
hdfs:NameNode:FSNamesystem:TransactionsSinceLastLogRoll dfs/FSNamesystem/TransactionsSinceLastLogRoll n
hdfs:NameNode:FSNamesystem:LastWrittenTransactionId dfs/FSNamesystem/LastWrittenTransactionId n
hdfs:NameNode:FSNamesystem:CapacityTotal dfs/FSNamesystem/CapacityTotal n
hdfs:NameNode:FSNamesystem:CapacityUsed dfs/FSNamesystem/CapacityUsed n
hdfs:NameNode:FSNamesystem:CapacityRemaining dfs/FSNamesystem/CapacityRemaining n
hdfs:NameNode:FSNamesystem:CapacityUsedNonDFS dfs/FSNamesystem/CapacityUsedNonDFS n
hdfs:NameNode:FSNamesystem:TotalLoad dfs/FSNamesystem/TotalLoad n
hdfs:NameNode:FSNamesystem:SnapshottableDirectories dfs/FSNamesystem/SnapshottableDirectories n
hdfs:NameNode:FSNamesystem:Snapshots dfs/FSNamesystem/Snapshots n
hdfs:NameNode:FSNamesystem:BlocksTotal dfs/FSNamesystem/BlocksTotal n
hdfs:NameNode:FSNamesystem:PendingReplicationBlocks dfs/FSNamesystem/PendingReplicationBlocks n
hdfs:NameNode:FSNamesystem:UnderReplicatedBlocks dfs/FSNamesystem/UnderReplicatedBlocks n
hdfs:NameNode:FSNamesystem:CorruptBlocks dfs/FSNamesystem/CorruptBlocks n
hdfs:NameNode:FSNamesystem:ScheduledReplicationBlocks dfs/FSNamesystem/ScheduledReplicationBlocks n
hdfs:NameNode:FSNamesystem:PendingDeletionBlocks dfs/FSNamesystem/PendingDeletionBlocks n
hdfs:NameNode:FSNamesystem:ExcessBlocks dfs/FSNamesystem/ExcessBlocks n
hdfs:NameNode:FSNamesystem:PostponedMisreplicatedBlocks dfs/FSNamesystem/PostponedMisreplicatedBlocks n
hdfs:NameNode:FSNamesystem:PendingDataNodeMessageCourt dfs/FSNamesystem/PendingDataNodeMessageCourt n
hdfs:NameNode:FSNamesystem:MillisSinceLastLoadedEdits dfs/FSNamesystem/MillisSinceLastLoadedEdits n
hdfs:NameNode:FSNamesystem:BlockCapacity dfs/FSNamesystem/BlockCapacity n
hdfs:NameNode:FSNamesystem:StaleDataNodes dfs/FSNamesystem/StaleDataNodes n
hdfs:NameNode:FSNamesystem:TotalFiles dfs/FSNamesystem/TotalFiles n
hdfs:NameNode:JvmMetrics:MemHeapUsedM dfs/jvm/MemHeapUsedM n
hdfs:NameNode:JvmMetrics:MemHeapCommittedM dfs/jvm/MemHeapCommittedM n
hdfs:NameNode:JvmMetrics:MemHeapMaxM dfs/jvm/MemHeapMaxM n
hdfs:NameNode:JvmMetrics:MemMaxM dfs/jvm/MemMaxM n

Métricas do YARN

Métrica Nome do Metrics Explorer Métricas ativadas
yarn:ResourceManager:ClusterMetrics:NumActiveNMs yarn/ClusterMetrics/NumActiveNMs y
yarn:ResourceManager:ClusterMetrics:NumDecommissionedNMs yarn/ClusterMetrics/NumDecommissionedNMs n
yarn:ResourceManager:ClusterMetrics:NumLostNMs yarn/ClusterMetrics/NumLostNMs n
yarn:ResourceManager:ClusterMetrics:NumUnhealthyNMs yarn/ClusterMetrics/NumUnhealthyNMs n
yarn:ResourceManager:ClusterMetrics:NumRebootedNMs yarn/ClusterMetrics/NumRebootedNMs n
yarn:ResourceManager:QueueMetrics:running_0 yarn/QueueMetrics/running_0 y
yarn:ResourceManager:QueueMetrics:running_60 yarn/QueueMetrics/running_60 y
yarn:ResourceManager:QueueMetrics:running_300 yarn/QueueMetrics/running_300 y
yarn:ResourceManager:QueueMetrics:running_1440 yarn/QueueMetrics/running_1440 y
yarn:ResourceManager:QueueMetrics:AppsSubmitted yarn/QueueMetrics/AppsSubmitted y
yarn:ResourceManager:QueueMetrics:AvailableMB yarn/QueueMetrics/AvailableMB y
yarn:ResourceManager:QueueMetrics:PendingContainers yarn/QueueMetrics/PendingContainers y
yarn:ResourceManager:QueueMetrics:AppsRunning yarn/QueueMetrics/AppsRunning n
yarn:ResourceManager:QueueMetrics:AppsPending yarn/QueueMetrics/AppsPending n
yarn:ResourceManager:QueueMetrics:AppsCompleted yarn/QueueMetrics/AppsCompleted n
yarn:ResourceManager:QueueMetrics:AppsKilled yarn/QueueMetrics/AppsKilled n
yarn:ResourceManager:QueueMetrics:AppsFailed yarn/QueueMetrics/AppsFailed n
yarn:ResourceManager:QueueMetrics:AllocatedMB yarn/QueueMetrics/AllocatedMB n
yarn:ResourceManager:QueueMetrics:AllocatedVCores yarn/QueueMetrics/AllocatedVCores n
yarn:ResourceManager:QueueMetrics:AllocatedContainers yarn/QueueMetrics/AllocatedContainers n
yarn:ResourceManager:QueueMetrics:AggregateContainersAllocated yarn/QueueMetrics/AggregateContainersAllocated n
yarn:ResourceManager:QueueMetrics:AggregateContainersReleased yarn/QueueMetrics/AggregateContainersReleased n
yarn:ResourceManager:QueueMetrics:AvailableVCores yarn/QueueMetrics/AvailableVCores n
yarn:ResourceManager:QueueMetrics:PendingMB yarn/QueueMetrics/PendingMB n
yarn:ResourceManager:QueueMetrics:PendingVCores yarn/QueueMetrics/PendingVCores n
yarn:ResourceManager:QueueMetrics:ReservedMB yarn/QueueMetrics/ReservedMB n
yarn:ResourceManager:QueueMetrics:ReservedVCores yarn/QueueMetrics/ReservedVCores n
yarn:ResourceManager:QueueMetrics:ReservedContainers yarn/QueueMetrics/ReservedContainers n
yarn:ResourceManager:QueueMetrics:ActiveUsers yarn/QueueMetrics/ActiveUsers n
yarn:ResourceManager:QueueMetrics:ActiveApplications yarn/QueueMetrics/ActiveApplications n
yarn:ResourceManager:QueueMetrics:FairShareMB yarn/QueueMetrics/FairShareMB n
yarn:ResourceManager:QueueMetrics:FairShareVCores yarn/QueueMetrics/FairShareVCores n
yarn:ResourceManager:QueueMetrics:MinShareMB yarn/QueueMetrics/MinShareMB n
yarn:ResourceManager:QueueMetrics:MinShareVCores yarn/QueueMetrics/MinShareVCores n
yarn:ResourceManager:QueueMetrics:MaxShareMB yarn/QueueMetrics/MaxShareMB n
yarn:ResourceManager:QueueMetrics:MaxShareVCores yarn/QueueMetrics/MaxShareVCores n
yarn:ResourceManager:JvmMetrics:MemHeapUsedM yarn/jvm/MemHeapUsedM n
yarn:ResourceManager:JvmMetrics:MemHeapCommittedM yarn/jvm/MemHeapCommittedM n
yarn:ResourceManager:JvmMetrics:MemHeapMaxM yarn/jvm/MemHeapMaxM n
yarn:ResourceManager:JvmMetrics:MemMaxM yarn/jvm/MemMaxM n

Métricas do Spark

Métricas do driver do Spark

Métrica Nome do Metrics Explorer Métricas ativadas
spark:driver:BlockManager:disk.diskSpaceUsed_MB spark/driver/BlockManager/disk/diskSpaceUsed_MB y
spark:driver:BlockManager:memory.maxMem_MB spark/driver/BlockManager/memory/maxMem_MB y
spark:driver:BlockManager:memory.memUsed_MB spark/driver/BlockManager/memory/memUsed_MB y
spark:driver:DAGScheduler:job.allJobs spark/driver/DAGScheduler/job/allJobs y
spark:driver:DAGScheduler:stage.failedStages spark/driver/DAGScheduler/stage/failedStages y
spark:driver:DAGScheduler:stage.waitingStages spark/driver/DAGScheduler/stage/waitingStages y

Métricas do executor do Spark

Métrica Nome do Metrics Explorer Métricas ativadas
spark:executor:executor:bytesRead spark/executor/bytesRead y
spark:executor:executor:bytesWritten spark/executor/bytesWritten y
spark:executor:executor:cpuTime spark/executor/cpuTime y
spark:executor:executor:diskBytesSpilled spark/executor/diskBytesSpilled y
spark:executor:executor:recordsRead spark/executor/recordsRead y
spark:executor:executor:recordsWritten spark/executor/recordsWritten y
spark:executor:executor:runTime spark/executor/runTime y
spark:executor:executor:shuffleRecordsRead spark/executor/shuffleRecordsRead y
spark:executor:executor:shuffleRecordsWritten spark/executor/shuffleRecordsWritten y
Métrica Nome do Metrics Explorer Métricas ativadas
flink:jobmanager:numRegisteredTaskManagers flink/jobmanager/numRegisteredTaskManagers n
flink:jobmanager:numRunningJobs flink/jobmanager/numRunningJobs n
flink:jobmanager:Status.JVM.ClassLoader.ClassesLoaded flink/jobmanager/Status.JVM.ClassLoader.ClassesLoaded n
flink:jobmanager:Status.JVM.ClassLoader.ClassesUnloaded flink/jobmanager/Status.JVM.ClassLoader.ClassesUnloaded n
flink:jobmanager:Status.JVM.CPU.Load flink/jobmanager/Status.JVM.CPU.Load n
flink:jobmanager:Status.JVM.CPU.Time flink/jobmanager/Status.JVM.CPU.Time y
flink:jobmanager:Status.JVM.GarbageCollector.PSMarkSweep.Count flink/jobmanager/Status.JVM.GarbageCollector.PSMarkSweep.Count n
flink:jobmanager:Status.JVM.GarbageCollector.PSMarkSweep.Time flink/jobmanager/Status.JVM.GarbageCollector.PSMarkSweep.Time n
flink:jobmanager:Status.JVM.GarbageCollector.PSScavenge.Count flink/jobmanager/Status.JVM.GarbageCollector.PSScavenge.Count n
flink:jobmanager:Status.JVM.GarbageCollector.PSScavenge.Time flink/jobmanager/Status.JVM.GarbageCollector.PSScavenge.Time n
flink:jobmanager:Status.JVM.Memory.Direct.Count flink/jobmanager/Status.JVM.Memory.Direct.Count y
flink:jobmanager:Status.JVM.Memory.Direct.MemoryUsed flink/jobmanager/Status.JVM.Memory.Direct.MemoryUsed y
flink:jobmanager:Status.JVM.Memory.Direct.TotalCapacity flink/jobmanager/Status.JVM.Memory.Direct.TotalCapacity y
flink:jobmanager:Status.JVM.Memory.Heap.Committed flink/jobmanager/Status.JVM.Memory.Heap.Committed y
flink:jobmanager:Status.JVM.Memory.Heap.Max flink/jobmanager/Status.JVM.Memory.Heap.Max y
flink:jobmanager:Status.JVM.Memory.Heap.Used flink/jobmanager/Status.JVM.Memory.Heap.Used y
flink:jobmanager:Status.JVM.Memory.Mapped.Count flink/jobmanager/Status.JVM.Memory.Mapped.Count y
flink:jobmanager:Status.JVM.Memory.Mapped.MemoryUsed flink/jobmanager/Status.JVM.Memory.Mapped.MemoryUsed y
flink:jobmanager:Status.JVM.Memory.Mapped.TotalCapacity flink/jobmanager/Status.JVM.Memory.Mapped.TotalCapacity y
flink:jobmanager:Status.JVM.Memory.Metaspace.Committed flink/jobmanager/Status.JVM.Memory.Metaspace.Committed n
flink:jobmanager:Status.JVM.Memory.Metaspace.Max flink/jobmanager/Status.JVM.Memory.Metaspace.Max n
flink:jobmanager:Status.JVM.Memory.Metaspace.Used flink/jobmanager/Status.JVM.Memory.Metaspace.Used n
flink:jobmanager:Status.JVM.Memory.NonHeap.Committed flink/jobmanager/Status.JVM.Memory.NonHeap.Committed n
flink:jobmanager:Status.JVM.Memory.NonHeap.Max flink/jobmanager/Status.JVM.Memory.NonHeap.Max n
flink:jobmanager:Status.JVM.Memory.NonHeap.Used flink/jobmanager/Status.JVM.Memory.NonHeap.Used n
flink:jobmanager:Status.JVM.Threads.Count flink/jobmanager/Status.JVM.Threads.Count n
flink:jobmanager:taskSlotsAvailable flink/jobmanager/taskSlotsAvailable y
flink:jobmanager:taskSlotsTotal flink/jobmanager/taskSlotsTotal y
flink:operator:numRecordsIn flink/operator/numRecordsIn n
flink:operator:numRecordsInPerSecond.count flink/operator/numRecordsInPerSecond.count n
flink:operator:numRecordsInPerSecond.rate flink/operator/numRecordsInPerSecond.rate n
flink:operator:numRecordsOut flink/operator/numRecordsOut n
flink:operator:numRecordsOutPerSecond.count flink/operator/numRecordsOutPerSecond.count n
flink:operator:numRecordsOutPerSecond.rate flink/operator/numRecordsOutPerSecond.rate n
flink:operator:numSplitsProcessed flink/operator/numSplitsProcessed n
flink:task:buffers.inPoolUsage flink/task/buffers.inPoolUsage n
flink:task:buffers.inputExclusiveBuffersUsage flink/task/buffers.inputExclusiveBuffersUsage n
flink:task:buffers.inputFloatingBuffersUsage flink/task/buffers.inputFloatingBuffersUsage n
flink:task:buffers.inputQueueLength flink/task/buffers.inputQueueLength n
flink:task:buffers.outPoolUsage flink/task/buffers.outPoolUsage n
flink:task:buffers.outputQueueLength flink/task/buffers.outputQueueLength n
flink:task:idleTimeMsPerSecond.count flink/task/idleTimeMsPerSecond.count n
flink:task:idleTimeMsPerSecond.rate flink/task/idleTimeMsPerSecond.rate n
flink:task:numBuffersInLocal flink/task/numBuffersInLocal n
flink:task:numBuffersInLocalPerSecond.count flink/task/numBuffersInLocalPerSecond.count n
flink:task:numBuffersInLocalPerSecond.rate flink/task/numBuffersInLocalPerSecond.rate n
flink:task:numBuffersInRemote flink/task/numBuffersInRemote n
flink:task:numBuffersInRemotePerSecond.count flink/task/numBuffersInRemotePerSecond.count n
flink:task:numBuffersInRemotePerSecond.rate flink/task/numBuffersInRemotePerSecond.rate n
flink:task:numBuffersOut flink/task/numBuffersOut n
flink:task:numBuffersOutPerSecond.count flink/task/numBuffersOutPerSecond.count n
flink:task:numBuffersOutPerSecond.rate flink/task/numBuffersOutPerSecond.rate n
flink:task:numBytesIn flink/task/numBytesIn n
flink:task:numBytesInLocal flink/task/numBytesInLocal n
flink:task:numBytesInLocalPerSecond.count flink/task/numBytesInLocalPerSecond.count n
flink:task:numBytesInLocalPerSecond.rate flink/task/numBytesInLocalPerSecond.rate n
flink:task:numBytesInPerSecond.count flink/task/numBytesInPerSecond.count n
flink:task:numBytesInPerSecond.rate flink/task/numBytesInPerSecond.rate n
flink:task:numBytesInRemote flink/task/numBytesInRemote n
flink:task:numBytesInRemotePerSecond.count flink/task/numBytesInRemotePerSecond.count n
flink:task:numBytesInRemotePerSecond.rate flink/task/numBytesInRemotePerSecond.rate n
flink:task:numBytesOut flink/task/numBytesOut n
flink:task:numBytesOutPerSecond.count flink/task/numBytesOutPerSecond.count n
flink:task:numBytesOutPerSecond.rate flink/task/numBytesOutPerSecond.rate n
flink:task:numRecordsIn flink/task/numRecordsIn n
flink:task:numRecordsInPerSecond.count flink/task/numRecordsInPerSecond.count n
flink:task:numRecordsInPerSecond.rate flink/task/numRecordsInPerSecond.rate n
flink:task:numRecordsOut flink/task/numRecordsOut n
flink:task:numRecordsOutPerSecond.count flink/task/numRecordsOutPerSecond.count n
flink:task:numRecordsOutPerSecond.rate flink/task/numRecordsOutPerSecond.rate n
flink:task:Shuffle.Netty.Input.Buffers.inPoolUsage flink/task/Shuffle.Netty.Input.Buffers.inPoolUsage n
flink:task:Shuffle.Netty.Input.Buffers.inputExclusiveBuffersUsage flink/task/Shuffle.Netty.Input.Buffers.inputExclusiveBuffersUsage n
flink:task:Shuffle.Netty.Input.Buffers.inputFloatingBuffersUsage flink/task/Shuffle.Netty.Input.Buffers.inputFloatingBuffersUsage n
flink:task:Shuffle.Netty.Input.Buffers.inputQueueLength flink/task/Shuffle.Netty.Input.Buffers.inputQueueLength n
flink:task:Shuffle.Netty.Input.numBuffersInLocal flink/task/Shuffle.Netty.Input.numBuffersInLocal n
flink:task:Shuffle.Netty.Input.numBuffersInLocalPerSecond.count flink/task/Shuffle.Netty.Input.numBuffersInLocalPerSecond.count n
flink:task:Shuffle.Netty.Input.numBuffersInLocalPerSecond.rate flink/task/Shuffle.Netty.Input.numBuffersInLocalPerSecond.rate n
flink:task:Shuffle.Netty.Input.numBuffersInRemote flink/task/Shuffle.Netty.Input.numBuffersInRemote n
flink:task:Shuffle.Netty.Input.numBuffersInRemotePerSecond.count flink/task/Shuffle.Netty.Input.numBuffersInRemotePerSecond.count n
flink:task:Shuffle.Netty.Input.numBuffersInRemotePerSecond.rate flink/task/Shuffle.Netty.Input.numBuffersInRemotePerSecond.rate n
flink:task:Shuffle.Netty.Input.numBytesInLocal flink/task/Shuffle.Netty.Input.numBytesInLocal n
flink:task:Shuffle.Netty.Input.numBytesInLocalPerSecond.count flink/task/Shuffle.Netty.Input.numBytesInLocalPerSecond.count n
flink:task:Shuffle.Netty.Input.numBytesInLocalPerSecond.rate flink/task/Shuffle.Netty.Input.numBytesInLocalPerSecond.rate n
flink:task:Shuffle.Netty.Input.numBytesInRemote flink/task/Shuffle.Netty.Input.numBytesInRemote n
flink:task:Shuffle.Netty.Input.numBytesInRemotePerSecond.count flink/task/Shuffle.Netty.Input.numBytesInRemotePerSecond.count n
flink:task:Shuffle.Netty.Input.numBytesInRemotePerSecond.rate flink/task/Shuffle.Netty.Input.numBytesInRemotePerSecond.rate n
flink:task:Shuffle.Netty.Output.Buffers.outPoolUsage flink/task/Shuffle.Netty.Output.Buffers.outPoolUsage n
flink:task:Shuffle.Netty.Output.Buffers.outputQueueLength flink/task/Shuffle.Netty.Output.Buffers.outputQueueLength n
flink:taskmanager:Status.flink.Memory.Managed.Total flink/taskmanager/Status.flink.Memory.Managed.Total n
flink:taskmanager:Status.flink.Memory.Managed.Used flink/taskmanager/Status.flink.Memory.Managed.Used n
flink:taskmanager:Status.JVM.ClassLoader.ClassesLoaded flink/taskmanager/Status.JVM.ClassLoader.ClassesLoaded n
flink:taskmanager:Status.JVM.ClassLoader.ClassesUnloaded flink/taskmanager/Status.JVM.ClassLoader.ClassesUnloaded n
flink:taskmanager:Status.JVM.CPU.Load flink/taskmanager/Status.JVM.CPU.Load n
flink:taskmanager:Status.JVM.CPU.Time flink/taskmanager/Status.JVM.CPU.Time y
flink:taskmanager:Status.JVM.GarbageCollector.PSMarkSweep.Count flink/taskmanager/Status.JVM.GarbageCollector.PSMarkSweep.Count n
flink:taskmanager:Status.JVM.GarbageCollector.PSMarkSweep.Time flink/taskmanager/Status.JVM.GarbageCollector.PSMarkSweep.Time n
flink:taskmanager:Status.JVM.GarbageCollector.PSScavenge.Count flink/taskmanager/Status.JVM.GarbageCollector.PSScavenge.Count n
flink:taskmanager:Status.JVM.GarbageCollector.PSScavenge.Time flink/taskmanager/Status.JVM.GarbageCollector.PSScavenge.Time n
flink:taskmanager:Status.JVM.Memory.Direct.Count flink/taskmanager/Status.JVM.Memory.Direct.Count y
flink:taskmanager:Status.JVM.Memory.Direct.MemoryUsed flink/taskmanager/Status.JVM.Memory.Direct.MemoryUsed y
flink:taskmanager:Status.JVM.Memory.Direct.TotalCapacity flink/taskmanager/Status.JVM.Memory.Direct.TotalCapacity y
flink:taskmanager:Status.JVM.Memory.Heap.Committed flink/taskmanager/Status.JVM.Memory.Heap.Committed y
flink:taskmanager:Status.JVM.Memory.Heap.Max flink/taskmanager/Status.JVM.Memory.Heap.Max y
flink:taskmanager:Status.JVM.Memory.Heap.Used flink/taskmanager/Status.JVM.Memory.Heap.Used y
flink:taskmanager:Status.JVM.Memory.Mapped.Count flink/taskmanager/Status.JVM.Memory.Mapped.Count y
flink:taskmanager:Status.JVM.Memory.Mapped.MemoryUsed flink/taskmanager/Status.JVM.Memory.Mapped.MemoryUsed y
flink:taskmanager:Status.JVM.Memory.Mapped.TotalCapacity flink/taskmanager/Status.JVM.Memory.Mapped.TotalCapacity y
flink:taskmanager:Status.JVM.Memory.Metaspace.Committed flink/taskmanager/Status.JVM.Memory.Metaspace.Committed n
flink:taskmanager:Status.JVM.Memory.Metaspace.Max flink/taskmanager/Status.JVM.Memory.Metaspace.Max n
flink:taskmanager:Status.JVM.Memory.Metaspace.Used flink/taskmanager/Status.JVM.Memory.Metaspace.Used n
flink:taskmanager:Status.JVM.Memory.NonHeap.Committed flink/taskmanager/Status.JVM.Memory.NonHeap.Committed n
flink:taskmanager:Status.JVM.Memory.NonHeap.Max flink/taskmanager/Status.JVM.Memory.NonHeap.Max n
flink:taskmanager:Status.JVM.Memory.NonHeap.Used flink/taskmanager/Status.JVM.Memory.NonHeap.Used n
flink:taskmanager:Status.JVM.Threads.Count flink/taskmanager/Status.JVM.Threads.Count n
flink:taskmanager:Status.Network.AvailableMemorySegments flink/taskmanager/Status.Network.AvailableMemorySegments n
flink:taskmanager:Status.Network.TotalMemorySegments flink/taskmanager/Status.Network.TotalMemorySegments n
flink:taskmanager:Status.Shuffle.Netty.AvailableMemory flink/taskmanager/Status.Shuffle.Netty.AvailableMemory n
flink:taskmanager:Status.Shuffle.Netty.AvailableMemorySegments flink/taskmanager/Status.Shuffle.Netty.AvailableMemorySegments n
flink:taskmanager:Status.Shuffle.Netty.TotalMemory flink/taskmanager/Status.Shuffle.Netty.TotalMemory n
flink:taskmanager:Status.Shuffle.Netty.TotalMemorySegments flink/taskmanager/Status.Shuffle.Netty.TotalMemorySegments n
flink:taskmanager:Status.Shuffle.Netty.UsedMemory flink/taskmanager/Status.Shuffle.Netty.UsedMemory n
flink:taskmanager:Status.Shuffle.Netty.UsedMemorySegments flink/taskmanager/Status.Shuffle.Netty.UsedMemorySegments n

Métricas do servidor de histórico do Spark

O Dataproc coleta as seguintes métricas de memória da JVM do serviço de histórico do Spark:

Métrica Nome do Metrics Explorer Métricas ativadas
sparkHistoryServer:JVM:Memory:HeapMemoryUsage.committed sparkHistoryServer/memory/CommittedHeapMemory y
sparkHistoryServer:JVM:Memory:HeapMemoryUsage.used sparkHistoryServer/memory/UsedHeapMemory y
sparkHistoryServer:JVM:Memory:HeapMemoryUsage.max sparkHistoryServer/memory/MaxHeapMemory y
sparkHistoryServer:JVM:Memory:NonHeapMemoryUsage.committed sparkHistoryServer/memory/CommittedNonHeapMemory y
sparkHistoryServer:JVM:Memory:NonHeapMemoryUsage.used sparkHistoryServer/memory/UsedNonHeapMemory y
sparkHistoryServer:JVM:Memory:NonHeapMemoryUsage.max sparkHistoryServer/memory/MaxNonHeapMemory y

Métricas do HiveServer 2

Métrica Nome do Metrics Explorer Métricas ativadas
hiveserver2:JVM:Memory:HeapMemoryUsage.committed hiveserver2/memory/CommittedHeapMemory y
hiveserver2:JVM:Memory:HeapMemoryUsage.used hiveserver2/memory/UsedHeapMemory y
hiveserver2:JVM:Memory:HeapMemoryUsage.max hiveserver2/memory/MaxHeapMemory y
hiveserver2:JVM:Memory:NonHeapMemoryUsage.committed hiveserver2/memory/CommittedNonHeapMemory y
hiveserver2:JVM:Memory:NonHeapMemoryUsage.used hiveserver2/memory/UsedNonHeapMemory y
hiveserver2:JVM:Memory:NonHeapMemoryUsage.max hiveserver2/memory/MaxNonHeapMemory y

Métricas do Hive Metastore

Métrica Nome do Metrics Explorer Métricas ativadas
hivemetastore:API:GetDatabase:Mean hivemetastore/get_database/mean y
hivemetastore:API:CreateDatabase:Mean hivemetastore/create_database/mean y
hivemetastore:API:DropDatabase:Mean hivemetastore/drop_database/mean y
hivemetastore:API:AlterDatabase:Mean hivemetastore/alter_database/mean y
hivemetastore:API:GetAllDatabases:Mean hivemetastore/get_all_databases/mean y
hivemetastore:API:CreateTable:Mean hivemetastore/create_table/mean y
hivemetastore:API:DropTable:Mean hivemetastore/drop_table/mean y
hivemetastore:API:AlterTable:Mean hivemetastore/alter_table/mean y
hivemetastore:API:GetTable:Mean hivemetastore/get_table/mean y
hivemetastore:API:GetAllTables:Mean hivemetastore/get_all_tables/mean y
hivemetastore:API:AddPartitionsReq:Mean hivemetastore/add_partitions_req/mean y
hivemetastore:API:DropPartition:Mean hivemetastore/drop_partition/mean y
hivemetastore:API:AlterPartition:Mean hivemetastore/alter_partition/mean y
hivemetastore:API:GetPartition:Mean hivemetastore/get_partition/mean y
hivemetastore:API:GetPartitionNames:Mean hivemetastore/get_partition_names/mean y
hivemetastore:API:GetPartitionsPs:Mean hivemetastore/get_partitions_ps/mean y
hivemetastore:API:GetPartitionsPsWithAuth:Mean hivemetastore/get_partitions_ps_with_auth/mean y

Medidas de métricas do Hive Metastore

Medida estatística Exemplo de métrica Exemplo de nome de métrica
Máx. hivemetastore:API:GetDatabase:Max hivemetastore/get_database/max
Mín. hivemetastore:API:GetDatabase:Min hivemetastore/get_database/min
Média hivemetastore:API:GetDatabase:Mean hivemetastore/get_database/mean
Contagem hivemetastore:API:GetDatabase:Count hivemetastore/get_database/count
50º percentil hivemetastore:API:GetDatabase:50thPercentile hivemetastore/get_database/median
75º percentil hivemetastore:API:GetDatabase:75thPercentile hivemetastore/get_database/75th_percentile
95thPercentile hivemetastore:API:GetDatabase:95thPercentile hivemetastore/get_database/95th_percentile
98º percentil hivemetastore:API:GetDatabase:98thPercentile hivemetastore/get_database/98th_percentile
99thPercentile hivemetastore:API:GetDatabase:99thPercentile hivemetastore/get_database/99th_percentile
999thPercentile hivemetastore:API:GetDatabase:999thPercentile hivemetastore/get_database/999th_percentile
StdDev hivemetastore:API:GetDatabase:StdDev hivemetastore/get_database/stddev
FifteenMinuteRate hivemetastore:API:GetDatabase:FifteenMinuteRate hivemetastore/get_database/15min_rate
FiveMinuteRate hivemetastore:API:GetDatabase:FiveMinuteRate hivemetastore/get_database/5min_rate
OneMinuteRate hivemetastore:API:GetDatabase:OneMinuteRate hivemetastore/get_database/1min_rate
MeanRate hivemetastore:API:GetDatabase:MeanRate hivemetastore/get_database/mean_rate

Métricas do agente de monitoramento do Dataproc

O Dataproc coleta as seguintes métricas do agente de monitoramento do Dataproc quando você define --metric-sources=monitoring-agent-defaults. Essas métricas são publicadas com o prefixo agent.googleapis.com.

CPU
agent.googleapis.com/cpu/load_15m
agent.googleapis.com/cpu/load_1m
agent.googleapis.com/cpu/load_5m
agent.googleapis.com/cpu/usage_time*
agent.googleapis.com/cpu/utilization*

Disco
agent.googleapis.com/disk/bytes_used
agent.googleapis.com/disk/io_time
agent.googleapis.com/disk/merged_operations
agent.googleapis.com/disk/operation_count
agent.googleapis.com/disk/operation_time
agent.googleapis.com/disk/pending_operations
agent.googleapis.com/disk/percent_used
agent.googleapis.com/disk/read_bytes_count

Troca
agent.googleapis.com/swap/bytes_used
agent.googleapis.com/swap/io
agent.googleapis.com/swap/percent_used

Memória
agent.googleapis.com/memory/bytes_used
agent.googleapis.com/memory/percent_used

Processos: (segue uma política de cotas ligeiramente diferente para alguns atributos)
agent.googleapis.com/processes/count_by_state
agent.googleapis.com/processes/cpu_time
agent.googleapis.com/processes/disk/read_bytes_count
agent.googleapis.com/processes/disk/write_bytes_count
agent.googleapis.com/processes/fork_count
agent.googleapis.com/processes/rss_usage
agent.googleapis.com/processes/vm_usage

Interface
agent.googleapis.com/interface/errors
agent.googleapis.com/interface/packets
agent.googleapis.com/interface/traffic

Rede
agent.googleapis.com/network/tcp_connections

Criar um painel de monitoramento

É possível criar um painel do Monitoring que mostra gráficos de métricas do Dataproc selecionadas.

  1. Selecione + CREATE DASHBOARD na página Dashboards Overview do Monitoring. Forneça um nome para o painel e clique em Add Chart no menu superior direito para abrir a janela Add Chart. Selecione “Cloud Dataproc Cluster” como o tipo de recurso. Selecione uma ou mais métricas e propriedades para métricas e gráficos. Em seguida, salve o gráfico.

  2. É possível adicionar gráficos ao seu painel. Depois que você salvar o painel, seu título aparecerá na página Dashboards Overview do Monitoring. Os gráficos do painel podem ser exibidos, atualizados e excluídos a partir da página de exibição do painel.

A seguir