{"componentChunkName":"component---src-templates-acg-portal-new-template-tsx","path":"/lmo7z6y11","result":{"data":{"markdownRemark":{"html":"

概述

\n

Milvus 引擎实例提供了通过控制台的方式查询和修改实例的配置信息。本文将介绍如何在控制台上更新 Milvus 引擎实例配置，以满足不同的业务需求。

\n

前提条件

\n\n

操作步骤

\n

按以下步骤在控制台上更新 Milvus 引擎实例的配置信息：

\n
    \n
  1. 登录云管理控制台，选择\"产品服务 > 数据库 > 向量数据库 VectorDB\"。
    2. \n
  2. 选择云服务器所在的区域。
    3. \n
  3. 在实例列表中找到对应的 Milvus 引擎实例，点击实例名称进入实例详情页。
    4. \n
  4. 在左侧导航栏选择\"实例配置\"页签。
    5. \n
  5. \n

    在\"实例配置\"输入框内输入需要覆盖默认配置的参数，然后点击\"保存配置\"。

    \n\n
    6. \n
  6. 在弹出的\"提示\"对话框中，输入变更原因，点击\"确定\"。
    7. \n
\n
\n

说明：配置修改请求提交后，若所修改配置项需要重启以生效，则在配置修改完成后将重启实例。此时实例将暂时进入升级中状态，待配置更新完成后，集群将自动恢复至运行中状态。

\n
\n\n
\n
\n
\n YAML\n \n
\n
\n 
1# Related configuration of rootCoord, used to handle data definition language (DDL) and data control language (DCL) requests\n2rootCoord:\n3  maxDatabaseNum: 64 # Maximum number of database\n4  maxPartitionNum: 4096 # Maximum number of partitions in a collection\n5  minSegmentSizeToEnableIndex: 1024 # It's a threshold. When the segment size is less than this value, the segment will not be indexed\n6  importTaskExpiration: 900 # (in seconds) Duration after which an import task will expire (be killed). Default 900 seconds (15 minutes).\n7  importTaskRetention: 86400 # (in seconds) Milvus will keep the record of import tasks for at least `importTaskRetention` seconds. Default 86400, seconds (24 hours).\n8  grpc:\n9    serverMaxSendSize: 536870912\n10    serverMaxRecvSize: 268435456\n11    clientMaxSendSize: 268435456\n12    clientMaxRecvSize: 536870912\n13    \n14# Related configuration of proxy, used to validate client requests and reduce the returned results.\n15proxy:\n16  timeTickInterval: 200 # ms, the interval that proxy synchronize the time tick\n17  healthCheckTimeout: 3000 # ms, the interval that to do component healthy check\n18  maxNameLength: 255 # Maximum length of name for a collection or alias\n19  # Maximum number of fields in a collection.\n20  # As of today (2.2.0 and after) it is strongly DISCOURAGED to set maxFieldNum >= 64.\n21  # So adjust at your risk!\n22  maxFieldNum: 64\n23  maxTaskNum: 1024 # max task number of proxy task queue\n24  grpc:\n25    serverMaxSendSize: 268435456\n26    serverMaxRecvSize: 67108864\n27    clientMaxSendSize: 268435456\n28    clientMaxRecvSize: 67108864\n29\n30# Related configuration of queryCoord, used to manage topology and load balancing for the query nodes, and handoff from growing segments to sealed segments.\n31queryCoord:\n32  autoHandoff: true # Enable auto handoff\n33  autoBalance: true # Enable auto balance\n34  balancer: ScoreBasedBalancer # Balancer to use\n35  overloadedMemoryThresholdPercentage: 90 # The threshold percentage that memory overload\n36  balanceIntervalSeconds: 60\n37  memoryUsageMaxDifferencePercentage: 30\n38  checkInterval: 1000\n39  channelTaskTimeout: 60000 # 1 minute\n40  segmentTaskTimeout: 120000 # 2 minute\n41  distPullInterval: 500\n42  heartbeatAvailableInterval: 10000 # 10s, Only QueryNodes which fetched heartbeats within the duration are available\n43  loadTimeoutSeconds: 600\n44  checkHandoffInterval: 5000\n45  grpc:\n46    serverMaxSendSize: 536870912\n47    serverMaxRecvSize: 268435456\n48    clientMaxSendSize: 268435456\n49    clientMaxRecvSize: 536870912\n50\n51# Related configuration of queryNode, used to run hybrid search between vector and scalar data.\n52queryNode:\n53  dataSync:\n54    flowGraph:\n55      maxQueueLength: 16 # Maximum length of task queue in flowgraph\n56      maxParallelism: 1024 # Maximum number of tasks executed in parallel in the flowgraph\n57  stats:\n58    publishInterval: 1000 # Interval for querynode to report node information (milliseconds)\n59  segcore:\n60    cgoPoolSizeRatio: 2.0 # cgo pool size ratio to max read concurrency\n61    knowhereThreadPoolNumRatio: 4\n62    # Use more threads to make better use of SSD throughput in disk index.\n63    # This parameter is only useful when enable-disk = true.\n64    # And this value should be a number greater than 1 and less than 32.\n65    chunkRows: 128 # The number of vectors in a chunk.\n66    exprEvalBatchSize: 8192 # The batch size for executor get next\n67    interimIndex: # build a vector temperate index for growing segment or binlog to accelerate search\n68      enableIndex: true\n69      nlist: 128 # segment index nlist\n70      nprobe: 16 # nprobe to search segment, based on your accuracy requirement, must smaller than nlist\n71      memExpansionRate: 1.15 # the ratio of building interim index memory usage to raw data\n72  loadMemoryUsageFactor: 1 # The multiply factor of calculating the memory usage while loading segments\n73  enableDisk: false # enable querynode load disk index, and search on disk index\n74  maxDiskUsagePercentage: 95\n75  grouping:\n76    enabled: true\n77    maxNQ: 1000\n78    topKMergeRatio: 20\n79  scheduler:\n80    receiveChanSize: 10240\n81    unsolvedQueueSize: 10240\n82    # maxReadConcurrentRatio is the concurrency ratio of read task (search task and query task).\n83    # Max read concurrency would be the value of runtime.NumCPU * maxReadConcurrentRatio.\n84    # It defaults to 2.0, which means max read concurrency would be the value of runtime.NumCPU * 2.\n85    # Max read concurrency must greater than or equal to 1, and less than or equal to runtime.NumCPU * 100.\n86    # (0, 100]\n87    maxReadConcurrentRatio: 1\n88    cpuRatio: 10 # ratio used to estimate read task cpu usage.\n89    maxTimestampLag: 86400\n90    # read task schedule policy: fifo(by default), user-task-polling.\n91    scheduleReadPolicy:\n92      # fifo: A FIFO queue support the schedule.\n93      # user-task-polling:\n94      #     The user's tasks will be polled one by one and scheduled.\n95      #     Scheduling is fair on task granularity.\n96      #     The policy is based on the username for authentication.\n97      #     And an empty username is considered the same user.\n98      #     When there are no multi-users, the policy decay into FIFO\n99      name: fifo\n100      maxPendingTask: 10240\n101      # user-task-polling configure:\n102      taskQueueExpire: 60 # 1 min by default, expire time of inner user task queue since queue is empty.\n103      enableCrossUserGrouping: false # false by default Enable Cross user grouping when using user-task-polling policy. (close it if task of any user can not merge others).\n104      maxPendingTaskPerUser: 1024 # 50 by default, max pending task in scheduler per user.\n105  grpc:\n106    serverMaxSendSize: 536870912\n107    serverMaxRecvSize: 268435456\n108    clientMaxSendSize: 268435456\n109    clientMaxRecvSize: 536870912\n110\n111indexCoord:\n112  bindIndexNodeMode:\n113    enable: false\n114    withCred: false\n115  segment:\n116    minSegmentNumRowsToEnableIndex: 1024 # It's a threshold. When the segment num rows is less than this value, the segment will not be indexed\n117\n118indexNode:\n119  scheduler:\n120    buildParallel: 1\n121  enableDisk: true # enable index node build disk vector index\n122  maxDiskUsagePercentage: 95\n123  grpc:\n124    serverMaxSendSize: 536870912\n125    serverMaxRecvSize: 268435456\n126    clientMaxSendSize: 268435456\n127    clientMaxRecvSize: 536870912\n128\n129dataCoord:\n130  channel:\n131    watchTimeoutInterval: 300 # Timeout on watching channels (in seconds). Datanode tickler update watch progress will reset timeout timer.\n132    balanceSilentDuration: 300 # The duration before the channelBalancer on datacoord to run\n133    balanceInterval: 360 #The interval for the channelBalancer on datacoord to check balance status\n134  segment:\n135    maxSize: 1024 # Maximum size of a segment in MB\n136    diskSegmentMaxSize: 2048 # Maximum size of a segment in MB for collection which has Disk index\n137    sealProportion: 0.12\n138    # The time of the assignment expiration in ms\n139    # Warning! this parameter is an expert variable and closely related to data integrity. Without specific\n140    # target and solid understanding of the scenarios, it should not be changed. If it's necessary to alter\n141    # this parameter, make sure that the newly changed value is larger than the previous value used before restart\n142    # otherwise there could be a large possibility of data loss\n143    assignmentExpiration: 2000\n144    maxLife: 86400 # The max lifetime of segment in seconds, 24*60*60\n145    # If a segment didn't accept dml records in maxIdleTime and the size of segment is greater than\n146    # minSizeFromIdleToSealed, Milvus will automatically seal it.\n147    # The max idle time of segment in seconds, 10*60.\n148    maxIdleTime: 600\n149    minSizeFromIdleToSealed: 16 # The min size in MB of segment which can be idle from sealed.\n150    # The max number of binlog file for one segment, the segment will be sealed if\n151    # the number of binlog file reaches to max value.\n152    maxBinlogFileNumber: 32\n153    smallProportion: 0.5 # The segment is considered as "small segment" when its # of rows is smaller than\n154    # (smallProportion * segment max # of rows).\n155    # A compaction will happen on small segments if the segment after compaction will have\n156    compactableProportion: 0.85\n157    # over (compactableProportion * segment max # of rows) rows.\n158    # MUST BE GREATER THAN OR EQUAL TO <smallProportion>!!!\n159    # During compaction, the size of segment # of rows is able to exceed segment max # of rows by (expansionRate-1) * 100%.\n160    expansionRate: 1.25\n161    # Whether to enable levelzero segment\n162    enableLevelZero: false\n163  enableCompaction: true # Enable data segment compaction\n164  compaction:\n165    enableAutoCompaction: true\n166    rpcTimeout: 10 # compaction rpc request timeout in seconds\n167    maxParallelTaskNum: 10 # max parallel compaction task number\n168    indexBasedCompaction: true\n169\n170    levelzero:\n171      forceTrigger:\n172        minSize: 8 # The minmum size in MB to force trigger a LevelZero Compaction\n173        deltalogMinNum: 10 # the minimum number of deltalog files to force trigger a LevelZero Compaction\n174\n175  enableGarbageCollection: true\n176  gc:\n177    interval: 3600 # gc interval in seconds\n178    missingTolerance: 3600 # file meta missing tolerance duration in seconds, 3600\n179    dropTolerance: 10800 # file belongs to dropped entity tolerance duration in seconds. 10800\n180  enableActiveStandby: false\n181  grpc:\n182    serverMaxSendSize: 536870912\n183    serverMaxRecvSize: 268435456\n184    clientMaxSendSize: 268435456\n185    clientMaxRecvSize: 536870912\n186\n187dataNode:\n188  dataSync:\n189    flowGraph:\n190      maxQueueLength: 16 # Maximum length of task queue in flowgraph\n191      maxParallelism: 1024 # Maximum number of tasks executed in parallel in the flowgraph\n192    maxParallelSyncMgrTasks: 256 #The max concurrent sync task number of datanode sync mgr globally \n193    skipMode:\n194      # when there are only timetick msg in flowgraph for a while (longer than coldTime),\n195      # flowGraph will turn on skip mode to skip most timeticks to reduce cost, especially there are a lot of channels\n196      enable: true\n197      skipNum: 4\n198      coldTime: 60\n199  segment:\n200    insertBufSize: 16777216 # Max buffer size to flush for a single segment.\n201    deleteBufBytes: 67108864 # Max buffer size to flush del for a single channel\n202    syncPeriod: 600 # The period to sync segments if buffer is not empty.\n203  # can specify ip for example\n204  # ip: 127.0.0.1\n205  grpc:\n206    serverMaxSendSize: 536870912\n207    serverMaxRecvSize: 268435456\n208    clientMaxSendSize: 268435456\n209    clientMaxRecvSize: 536870912\n210  memory:\n211    forceSyncEnable: true # `true` to force sync if memory usage is too high\n212    forceSyncSegmentNum: 1 # number of segments to sync, segments with top largest buffer will be synced.\n213    watermarkStandalone: 0.2 # memory watermark for standalone, upon reaching this watermark, segments will be synced.\n214    watermarkCluster: 0.5 # memory watermark for cluster, upon reaching this watermark, segments will be synced.\n215  timetick:\n216    byRPC: true\n217  channel:\n218    # specify the size of global work pool of all channels\n219    # if this parameter <= 0, will set it as the maximum number of CPUs that can be executing\n220    # suggest to set it bigger on large collection numbers to avoid blocking\n221    workPoolSize: -1\n222    # specify the size of global work pool for channel checkpoint updating\n223    # if this parameter <= 0, will set it as 1000\n224    # suggest to set it bigger on large collection numbers to avoid blocking\n225    updateChannelCheckpointMaxParallel: 1000\n226\n227grpc:\n228  client:\n229    compressionEnabled: false\n230    dialTimeout: 200\n231    keepAliveTime: 10000\n232    keepAliveTimeout: 20000\n233    maxMaxAttempts: 10\n234    initialBackOff: 0.2 # seconds\n235    maxBackoff: 10 # seconds\n236    \n237quotaAndLimits:\n238  enabled: true # `true` to enable quota and limits, `false` to disable.\n239  limits:\n240    maxCollectionNum: 65536\n241    maxCollectionNumPerDB: 65536\n242  # quotaCenterCollectInterval is the time interval that quotaCenter\n243  # collects metrics from Proxies, Query cluster and Data cluster.\n244  # seconds, (0 ~ 65536)\n245  quotaCenterCollectInterval: 3\n246  ddl:\n247    enabled: false\n248    collectionRate: -1 # qps, default no limit, rate for CreateCollection, DropCollection, LoadCollection, ReleaseCollection\n249    partitionRate: -1 # qps, default no limit, rate for CreatePartition, DropPartition, LoadPartition, ReleasePartition\n250  indexRate:\n251    enabled: false\n252    max: -1 # qps, default no limit, rate for CreateIndex, DropIndex\n253  flushRate:\n254    enabled: false\n255    max: -1 # qps, default no limit, rate for flush\n256  compactionRate:\n257    enabled: false\n258    max: -1 # qps, default no limit, rate for manualCompaction\n259  dml:\n260    # dml limit rates, default no limit.\n261    # The maximum rate will not be greater than max.\n262    enabled: false\n263    insertRate:\n264      collection:\n265        max: -1 # MB/s, default no limit\n266      max: -1 # MB/s, default no limit\n267    upsertRate:\n268      collection:\n269        max: -1 # MB/s, default no limit\n270      max: -1 # MB/s, default no limit\n271    deleteRate:\n272      collection:\n273        max: -1 # MB/s, default no limit\n274      max: -1 # MB/s, default no limit\n275    bulkLoadRate:\n276      collection:\n277        max: -1 # MB/s, default no limit, not support yet. TODO: limit bulkLoad rate\n278      max: -1 # MB/s, default no limit, not support yet. TODO: limit bulkLoad rate\n279  dql:\n280    # dql limit rates, default no limit.\n281    # The maximum rate will not be greater than max.\n282    enabled: false\n283    searchRate:\n284      collection:\n285        max: -1 # vps (vectors per second), default no limit\n286      max: -1 # vps (vectors per second), default no limit\n287    queryRate:\n288      collection:\n289        max: -1 # qps, default no limit\n290      max: -1 # qps, default no limit\n291  limitWriting:\n292    # forceDeny false means dml requests are allowed (except for some\n293    # specific conditions, such as memory of nodes to water marker), true means always reject all dml requests.\n294    forceDeny: false\n295    ttProtection:\n296      enabled: false\n297      # maxTimeTickDelay indicates the backpressure for DML Operations.\n298      # DML rates would be reduced according to the ratio of time tick delay to maxTimeTickDelay,\n299      # if time tick delay is greater than maxTimeTickDelay, all DML requests would be rejected.\n300      # seconds\n301      maxTimeTickDelay: 300\n302    memProtection:\n303      # When memory usage > memoryHighWaterLevel, all dml requests would be rejected;\n304      # When memoryLowWaterLevel < memory usage < memoryHighWaterLevel, reduce the dml rate;\n305      # When memory usage < memoryLowWaterLevel, no action.\n306      enabled: true\n307      dataNodeMemoryLowWaterLevel: 0.85 # (0, 1], memoryLowWaterLevel in DataNodes\n308      dataNodeMemoryHighWaterLevel: 0.95 # (0, 1], memoryHighWaterLevel in DataNodes\n309      queryNodeMemoryLowWaterLevel: 0.85 # (0, 1], memoryLowWaterLevel in QueryNodes\n310      queryNodeMemoryHighWaterLevel: 0.95 # (0, 1], memoryHighWaterLevel in QueryNodes\n311    growingSegmentsSizeProtection:\n312      # No action will be taken if the growing segments size is less than the low watermark.\n313      # When the growing segments size exceeds the low watermark, the dml rate will be reduced,\n314      # but the rate will not be lower than `minRateRatio * dmlRate`.\n315      enabled: false\n316      minRateRatio: 0.5\n317      lowWaterLevel: 0.2\n318      highWaterLevel: 0.4\n319    diskProtection:\n320      enabled: true # When the total file size of object storage is greater than `diskQuota`, all dml requests would be rejected;\n321      diskQuota: -1 # MB, (0, +inf), default no limit\n322      diskQuotaPerCollection: -1 # MB, (0, +inf), default no limit\n323  limitReading:\n324    # forceDeny false means dql requests are allowed (except for some\n325    # specific conditions, such as collection has been dropped), true means always reject all dql requests.\n326    forceDeny: false\n327    queueProtection:\n328      enabled: false\n329      # nqInQueueThreshold indicated that the system was under backpressure for Search/Query path.\n330      # If NQ in any QueryNode's queue is greater than nqInQueueThreshold, search&query rates would gradually cool off\n331      # until the NQ in queue no longer exceeds nqInQueueThreshold. We think of the NQ of query request as 1.\n332      # int, default no limit\n333      nqInQueueThreshold: -1\n334      # queueLatencyThreshold indicated that the system was under backpressure for Search/Query path.\n335      # If dql latency of queuing is greater than queueLatencyThreshold, search&query rates would gradually cool off\n336      # until the latency of queuing no longer exceeds queueLatencyThreshold.\n337      # The latency here refers to the averaged latency over a period of time.\n338      # milliseconds, default no limit\n339      queueLatencyThreshold: -1\n340    resultProtection:\n341      enabled: false\n342      # maxReadResultRate indicated that the system was under backpressure for Search/Query path.\n343      # If dql result rate is greater than maxReadResultRate, search&query rates would gradually cool off\n344      # until the read result rate no longer exceeds maxReadResultRate.\n345      # MB/s, default no limit\n346      maxReadResultRate: -1\n347    # colOffSpeed is the speed of search&query rates cool off.\n348    # (0, 1]\n349    coolOffSpeed: 0.9
\n
\n
\n
\n ","fields":{"slug":"lmo7z6y11","title":"修改实例配置","date":"2026-04-21","extractedHeadings":[]},"headings":[{"value":"概述","depth":2},{"value":"前提条件","depth":2},{"value":"操作步骤","depth":2}]}},"pageContext":{"isCreatedByStatefulCreatePages":false,"slug":"lmo7z6y11","prev":{"id":"tmnod5bo6","name":"配置变更","path":"tmnod5bo6","filePath":"操作指南/Milvus引擎/实例管理/配置变更.md","seo":null,"parentIds":["Plr8wkd1f","Umnochq1a","dmnod5bl1"],"parents":[{"id":"Plr8wkd1f","documentId":"49acb227-91bb-4cc7-b257-41f76f309b3c","name":"操作指南","repoName":"VDB","filePath":"操作指南","disabled":false,"path":"Plr8wkd1f","lastMergeTime":null,"isApiDoc":null,"httpMethod":null,"seo":null,"sourceOrgName":null,"sourceRepoName":null,"sourceDocumentId":null},{"id":"Umnochq1a","documentId":"904a850e-fc1e-4cf8-b6fb-7fa54f1e2a76","name":"Milvus引擎","repoName":"VDB","filePath":"操作指南/Milvus引擎","disabled":false,"path":"Umnochq1a","lastMergeTime":null,"isApiDoc":null,"httpMethod":null,"seo":null,"sourceOrgName":null,"sourceRepoName":null,"sourceDocumentId":""},{"id":"dmnod5bl1","documentId":"d3bef9dd-7b06-4f81-9cbe-48852d43a547","name":"实例管理","repoName":"VDB","filePath":"操作指南/Milvus引擎/实例管理","disabled":false,"path":"dmnod5bl1","lastMergeTime":null,"isApiDoc":null,"httpMethod":null,"seo":null,"sourceOrgName":null,"sourceRepoName":null,"sourceDocumentId":""}]},"next":{"id":"fmnod5bw7","name":"实例监控","path":"fmnod5bw7","filePath":"操作指南/Milvus引擎/实例监控/查看监控.md","seo":null,"parentIds":["Plr8wkd1f","Umnochq1a","5mnod5bvl"],"parents":[{"id":"Plr8wkd1f","documentId":"49acb227-91bb-4cc7-b257-41f76f309b3c","name":"操作指南","repoName":"VDB","filePath":"操作指南","disabled":false,"path":"Plr8wkd1f","lastMergeTime":null,"isApiDoc":null,"httpMethod":null,"seo":null,"sourceOrgName":null,"sourceRepoName":null,"sourceDocumentId":null},{"id":"Umnochq1a","documentId":"904a850e-fc1e-4cf8-b6fb-7fa54f1e2a76","name":"Milvus引擎","repoName":"VDB","filePath":"操作指南/Milvus引擎","disabled":false,"path":"Umnochq1a","lastMergeTime":null,"isApiDoc":null,"httpMethod":null,"seo":null,"sourceOrgName":null,"sourceRepoName":null,"sourceDocumentId":""},{"id":"5mnod5bvl","documentId":"558ffdf5-acd5-4f2b-ad90-907f5c9e14eb","name":"实例监控","repoName":"VDB","filePath":"操作指南/Milvus引擎/实例监控","disabled":false,"path":"5mnod5bvl","lastMergeTime":null,"isApiDoc":null,"httpMethod":null,"seo":null,"sourceOrgName":null,"sourceRepoName":null,"sourceDocumentId":""}]},"parents":[{"id":"Plr8wkd1f","documentId":"49acb227-91bb-4cc7-b257-41f76f309b3c","name":"操作指南","repoName":"VDB","filePath":"操作指南","disabled":false,"path":"Plr8wkd1f","lastMergeTime":null,"isApiDoc":null,"httpMethod":null,"seo":null,"sourceOrgName":null,"sourceRepoName":null,"sourceDocumentId":null},{"id":"Umnochq1a","documentId":"904a850e-fc1e-4cf8-b6fb-7fa54f1e2a76","name":"Milvus引擎","repoName":"VDB","filePath":"操作指南/Milvus引擎","disabled":false,"path":"Umnochq1a","lastMergeTime":null,"isApiDoc":null,"httpMethod":null,"seo":null,"sourceOrgName":null,"sourceRepoName":null,"sourceDocumentId":""},{"id":"dmnod5bl1","documentId":"d3bef9dd-7b06-4f81-9cbe-48852d43a547","name":"实例管理","repoName":"VDB","filePath":"操作指南/Milvus引擎/实例管理","disabled":false,"path":"dmnod5bl1","lastMergeTime":null,"isApiDoc":null,"httpMethod":null,"seo":null,"sourceOrgName":null,"sourceRepoName":null,"sourceDocumentId":""}],"specificSeo":null}}}