Hadoop指标
Download as TXT, PDF0 likes219 views
This document lists various Hadoop metrics for the Datanode, JVM, RPC, and MapReduce components. It includes metrics such as the number of blocks read/written, average time for block operations, JVM memory and thread usage, number of open RPC connections, and MapReduce shuffle and task metrics. The metrics provide insight into the performance and resource usage of the different Hadoop services and components.
Hadoop指标
- 1. hadoop¸÷Ïî¿ØÖÆÖ¸±ê˵Ã÷£º hadoop-hdfs-datanode dfs.datanode.blockChecksumOp_avg_time ¿éУÑéƽ¾ùʱ¼ä dfs.datanode.blockChecksumOp_num_ops ¿é¼ìÑé´ÎÊý dfs.datanode.blockReports_avg_time ¿é±¨¸æƽ¾ùʱ¼ä dfs.datanode.blockReports_num_ops ¿é±¨¸æ´ÎÊý dfs.datanode.block_verification_failures ¿éÑé֤ʧ°Ü´ÎÊý dfs.datanode.blocks_read ´ÓÓ²Å̶Á¿é×Ü´ÎÊý dfs.datanode.blocks_removed ɾ³ý¿éÊýÄ¿ dfs.datanode.blocks_replicated ¿é¸´ÖÆ×Ü´ÎÊý dfs.datanode.blocks_verified ¿éÑéÖ¤×Ü´ÎÊý dfs.datanode.blocks_written ÏòÓ²ÅÌд¿é×Ü´ÎÊý dfs.datanode.bytes_read ¶Á³ö×Ü×Ö½Ú°üº¬crcÑéÖ¤Îļþ×Ö½ÚÊý dfs.datanode.bytes_written дÈë×Ü×Ö½ÚÊý£¨ÔÚÐ ´Èëÿ¸öpacketʱ¼ÆÊý£© dfs.datanode.copyBlockOp_avg_time ¸´ÖÆ¿éƽ¾ùʱ¼ä £¨µ¥Î»ms£© dfs.datanode.copyBlockOp_num_ops ¸´ÖÆ¿é´ÎÊý dfs.datanode.heartBeats_avg_time Ïònamenode»ã±¨Æ½¾ùʱ¼ä dfs.datanode.heartBeats_num_ops Ïònamenode»ã±¨×Ü´ÎÊý dfs.datanode.readBlockOp_avg_time ¶Á¿éƽ¾ùʱ¼ä£¨µ¥Î»ms£© dfs.datanode.readBlockOp_num_ops ¶Á¿é×Ü´ÎÊý Ò»°ãºÍdfs.datanode.blocks_read Ò»Ö£¬ÏÈ ´ÓÓ²Å̶ÁÈëÊäÈëÁ÷£¬Ôö¼Ódfs.datanode.blocks_read ¼ÆÊý£¬È»ºóÔÙÔö¼Ó¸Ã¼ÆÊý dfs.datanode.reads_from_local_client ´Ó±¾µØ¶ÁÈë¿é´ÎÊý dfs.datanode.reads_from_remote_client ´ÓÔ¶³Ì¶ÁÈë¿é´ÎÊý dfs.datanode.replaceBlockOp_avg_time Ìæ»»¿éƽ¾ùʱ¼ä£¨¸ºÔؾùºâ²ßÂÔ£© dfs.datanode.replaceBlockOp_num_ops Ìæ»»¿é´ÎÊý£¨¸ºÔؾùºâ²ßÂÔ£© dfs.datanode.volumeFailures notfound ºÍblockÓµÓеÄvolume ʧ°ÜÓÐ¹Ø dfs.datanode.writeBlockOp_avg_time д¿éƽ¾ùʱ¼ä dfs.datanode.writeBlockOp_num_ops д¿é×Ü ´ÎÊýÒ»°ãºÍdfs.datanode.blocks_written Ò»Ö£¬ÏÈ ´ÓÓ²ÅÌ£¬Ôö¼Ódfs.datanode.blocks_read ¼ÆÊý£¬È»ºóÔÙÔö¼Ó¸Ã¼ÆÊý dfs.datanode.writes_from_local_client д±¾µØ´ÎÊý dfs.datanode.writes_from_remote_client дԶ³Ì´ÎÊý jvm.metrics.gcCount gc×Ü´ÎÊý jvm.metrics.gcTimeMillis gc×ܺÄʱ(ms) jvm.metrics.logError jvm error ´ÎÊý jvm.metrics.logFatal jvm³öÏÖfatal´ÎÊý jvm.metrics.logInfo jvm info³öÏÖ´ÎÊý jvm.metrics.logWarn jvm warn³öÏÖ´ÎÊý jvm.metrics.maxMemoryM jvmÊÔͼʹÓÃ×î´óÄÚ ´æ£¨M£©£¬Èç¹ûûÓÐÏÞÖÆ·µ»ØLong.MAX_VALUE jvm.metrics.memHeapCommittedM jvmÌá½»¶ÑÄÚ´æ´óС jvm.metrics.memHeapUsedM jvmʹÓöÑÄÚ´æ´óС jvm.metrics.memNonHeapCommittedM jvm·Ç¶ÑÄÚ´æÒÑÌá½»´óС jvm.metrics.memNonHeapUsedM jvm·Ç¶ÑÄÚ´æÒÑʹÓôóС jvm.metrics.threadsBlocked ÕýÔÚ×èÈûµÈ´ý¼àÊÓÆ÷ËøµÄÏß³ÌÊýÄ¿ jvm.metrics.threadsNew ÉÐδÆô¶¯µÄÏß³ÌÊýÄ¿ jvm.metrics.threadsRunnable ÕýÔÚÖ´ÐÐ״̬µÄÏß³ÌÊýÄ¿ jvm.metrics.threadsTerminated ÒÑÍ˳öÏß³ÌÊýÄ¿ jvm.metrics.threadsTimedWaiting µÈ´ýÁíÒ»¸öÏß³ÌÖ ´ÐÐÈ¡¾öÓÚÖ¸¶¨µÈ´ýʱ¼äµÄ²Ù×÷µÄÏß³ÌÊýÄ¿ jvm.metrics.threadsWaiting #ÎÞÏÞÆڵصȴýÁíÒ»¸öÏß³ÌÀ´Ö ´ÐÐijһÌض¨²Ù×÷µÄÏß³ÌÊýÄ¿ rpc.metrics.NumOpenConnections #number of open connections rpcÁ¬½Ó´ò¿ªµÄÊýÄ¿ rpc.metrics.ReceivedBytes #number of bytes received rpcÊÕµ½µÄ×Ö½ÚÊý
- 2. rpc.metrics.RpcProcessingTime_avg_time #Average time for RPC Operations in last interval rpcÔÚ×î½üµÄ½»»¥ÖÐƽ¾ù²Ù×÷ʱ¼ä rpc.metrics.RpcProcessingTime_num_ops rpcÔÚ×î½üµÄ½»»¥ÖÐÁ¬½ÓÊýÄ¿ rpc.metrics.RpcQueueTime_avg_time rpcÔÚ½»»¥ÖÐƽ¾ùµÈ´ýʱ¼ä rpc.metrics.RpcQueueTime_num_ops rpc queueÖÐÍê³ÉµÄrpc²Ù×÷ÊýÄ¿ rpc.metrics.SentBytes number of bytes sent rpc· ¢Ë͵ÄÊý¾Ý×Ö½Ú rpc.metrics.callQueueLen length of the rpc queue rpc ¶ÓÁг¤¶È rpc.metrics.rpcAuthenticationFailures #number of failed authentications rpc Ñé֤ʧ°Ü´ÎÊý rpc.metrics.rpcAuthenticationSuccesses #number of successful authentications ÑéÖ¤³É¹¦Êý rpc.metrics.rpcAuthorizationFailures #number of failed authorizations ÊÚȨʧ°Ü´ÎÊý rpc.metrics.rpcAuthorizationSuccesses #number of successful authorizations ³É¹¦´ÎÊý mapred.shuffleInput.shuffle_failed_fetches # ´ÓmapÊä³öÖÐÈ¡Êý¾Ý¹ý³ÌÖлñȡʧ°Ü´ÎÊý mapred.shuffleInput.shuffle_fetchers_busy_percent #ÔÚ»ñÈ¡mapÊä³ö¹ý³ÌÖв ¢ÐлñÈ¡Ïß³Ìæµռ×ܲ¢ÐлñÈ¡Ïß³Ì°Ù·Ö±È mapred.shuffleInput.shuffle_input_bytes #shuffle¹ý³ÌÖжÁÈëÊý¾Ý×Ö½Ú mapred.shuffleInput.shuffle_success_fetches # ´ÓmapÊä³öÖÐÈ¡Êý¾Ý¹ý³ÌÖлñÈ¡³É¹¦´ÎÊý mapred.shuffleOutput.shuffle_failed_outputs #Ïòreduce·¢ËÍmapÊä³öʧ°Ü ´ÎÊý mapred.shuffleOutput.shuffle_handler_busy_percent #Ïòreduce· ¢ËÍmapÊä³öÖÐserverÏß³Ìæµռ×ܹ¤×÷Ïß³Ì(ÔÚtasktracker.http.threadsÖÐÅäÖÃ)°Ù·Ö±È¡£ mapred.shuffleOutput.shuffle_output_bytes #shuffle¹ý³ÌÖÐÊä³öÊý¾Ý×Ö½Ú mapred.shuffleOutput.shuffle_success_outputs Ïòreduce³É¹¦ mapred.tasktracker.mapTaskSlots ÉèÖÃmap²ÛÊý mapred.tasktracker.maps_running ÕýÔÚÔËÐеÄmapÊý mapred.tasktracker.reduceTaskSlots ÉèÖÃreduce²ÛÊý mapred.tasktracker.reduces_running ÕýÔÚÔËÐеÄreduceÊý mapred.tasktracker.tasks_completed Íê³ÉÈÎÎñÊý mapred.tasktracker.tasks_failed_ping #ÒòtasktrackerÓëtask½»»¥Ê§°Üµ¼ÖµÄʧ°ÜµÄtaskÊýÄ¿ mapred.tasktracker.tasks_failed_timeout #ÒòtaskÎ ´ÔÚmapred.task.timeout ÅäÖõÄ(ĬÈÏ10·ÖÖÓ)ʱ¼äÄڻ㱨½ø¶È¶ø³¬Ê±killµÄtaskÊýÄ¿ rpc.detailed-metrics.canCommit_avg_time rpcѯÎÊÊÇ·ñÌá½»ÈÎÎñƽ¾ùʱ¼ä rpc.detailed-metrics.canCommit_num_ops rpcѯÎÊÊÇ·ñÌá½»ÈÎÎñ´ÎÊý rpc.detailed-metrics.commitPending_avg_time #rpc±¨¸æÈÎÎñÌá½»Íê³É£¬µ«ÊǸÃÌá½»ÈÔÈ»´¦ÓÚpending״̬µÄƽ¾ùʱ¼ä rpc.detailed-metrics.commitPending_num_ops #rpc±¨¸æÈÎÎñÌá½»Íê³É£¬µ«ÊǸÃÌá½»ÈÔÈ»´¦ÓÚpending״̬µÄ´ÎÊý rpc.detailed-metrics.done_avg_time rpc±¨¸æÈÎÎñ³É¹¦Íê³ÉµÄƽ¾ùʱ¼ä rpc.detailed-metrics.done_num_ops rpc±¨¸æÈÎÎñ³É¹¦Íê³ÉµÄ´ÎÊý rpc.detailed-metrics.fatalError_avg_time rpc±¨¸æÈÎÎñ³öÏÖfatalerrorµÄƽ¾ùʱ¼ä rpc.detailed-metrics.fatalError_num_ops rpc±¨¸æÈÎÎñ³öÏÖfatalerrorµÄ ´ÎÊý rpc.detailed-metrics.getBlockInfo_avg_time ´ÓÖ¸¶¨datanode»ñÈ¡blockµÄƽ¾ùʱ¼ä rpc.detailed-metrics.getBlockInfo_num_ops ´ÓÖ¸¶¨datanode»ñÈ¡blockµÄ´ÎÊý rpc.detailed-metrics.getMapCompletionEvents_avg_time #reduce»ñÈ¡ÒѾ- Íê³ÉµÄmapÊä³öµØַʼþµÄƽ¾ùʱ¼ä rpc.detailed-metrics.getMapCompletionEvents_num_ops #reduce»ñÈ¡ÒѾ-
- 3. Íê³ÉµÄmapÊä³öµØַʼþµÄ´ÎÊý rpc.detailed-metrics.getProtocolVersion_avg_time #»ñÈ¡rpcÐÒé°æ±¾ÐÅÏ ¢µÄƽ¾ùʱ¼ä rpc.detailed-metrics.getProtocolVersion_num_ops #»ñÈ¡rpcÐÒé°æ±¾ÐÅÏ¢µÄ´ÎÊý rpc.detailed-metrics.getTask_avg_time µ±×Ó½ø³ÌÆô¶¯ºó£¬»ñÈ¡jvmtaskµÄƽ¾ùʱ¼ä rpc.detailed-metrics.getTask_num_ops µ±×Ó½ø³ÌÆô¶¯ºó£¬»ñÈ¡jvmtaskµÄ´ÎÊý rpc.detailed-metrics.ping_avg_time ×Ó½ø³ÌÖÜÆÚÐԵļì²â¸¸½ø³ÌÊÇ·ñ»¹ ´æ»îµÄƽ¾ùʱ¼ä rpc.detailed-metrics.ping_num_ops ×Ó½ø³ÌÖÜÆÚÐԵļì²â¸¸½ø³ÌÊÇ·ñ»¹ ´æ»îµÄ´ÎÊý rpc.detailed-metrics.recoverBlock_avg_time #Ϊָ¶¨µÄblock¿ªÊ¼»Ö¸ ´±ê¼ÇÉú³ÉµÄƽ¾ùʱ¼ä rpc.detailed-metrics.recoverBlock_num_ops Ϊָ¶¨µÄblock¿ªÊ¼»Ö¸´±ê¼ÇÉú³ÉµÄ ´ÎÊý rpc.detailed-metrics.reportDiagnosticInfo_avg_time #Ïò¸¸½ø³Ì±¨¸æÈÎÎñ´íÎóÏûÏ ¢µÄƽ¾ùʱ¼ä£¬¸Ã²Ù×÷Ó¦¾¡¿ÉÄÜÉÙ£¬ÕâЩÏûÏ¢»áÔÚjobtrackerÖб£´æ rpc.detailed-metrics.reportDiagnosticInfo_num_ops #Ïò¸¸½ø³Ì±¨¸æÈÎÎñ´íÎóÏûÏ ¢µÄ´ÎÊý rpc.detailed-metrics.startBlockRecovery_avg_time ¿ªÊ¼»Ö¸´blockµÄƽ¾ùʱ¼ä rpc.detailed-metrics.startBlockRecovery_num_ops ¿ªÊ¼»Ö¸´blockµÄ´ÎÊý rpc.detailed-metrics.statusUpdate_avg_time »ã±¨×Ó½ø³Ì½ø¶È¸ø¸¸½ø³ÌµÄƽ¾ùʱ¼ä rpc.detailed-metrics.statusUpdate_num_ops »ã±¨×Ó½ø³Ì½ø¶È¸ø¸¸½ø³ÌµÄ´ÎÊý rpc.detailed-metrics.updateBlock_avg_time #¸üÐÂblockµ½Ðµıê¼Ç¼°³¤¶ÈµÄƽ¾ù²Ù×÷ʱ¼ä rpc.detailed-metrics.updateBlock_num_ops #¸üÐÂblockµ½Ðµıê¼Ç¼°³¤¶ÈµÄ´ÎÊý