zoukankan      html  css  js  c++  java
  • MMS关键指标意义&各数值区间意义

    MMS关键指标意义&各数值区间意义


    What's MMS

    MongoDB Management Service (MMS) is a suite of services for managing MongoDB deployments.

    统计图表的数据来源

    all statistics can show in mongo shell by:

    >db.serverStatus()
    

    1. opcounters
      意义:The average number of commands performed per second since the last data point
      代表忙不忙,越小越好。(看业务)
      opcounters
    2. queue

    意义:The number of operations queued waiting for any lock -- 排队未执行的命令
    越小越好,平时<1
    此处输入图片的描述
    3. connection

    意义:The number of currently active connections to this server. A stack is allocated per connection; thus very many connections can result in significant RAM usage.
    越小越好。
    =client (connected) * 集群server 数目
    此处输入图片的描述
    4. lock

    意义:The percent of time write locked. The effective lock % is the percent of time in the global lock plus the percent of time locked by the hottest database. Because the data is sampled and combined, it is possible to see values over 100%.
    越小越好。
    造成锁增加的行为:update delete insert ...
    此处输入图片的描述
    5. Btree

    意义:A large number of misses means that you are indexes are too big to fit in RAM, which can cause a significant performance penalty -- 因为数据不再mem,从disk 加载到mem。
    mongodb 使用 btree index
    missratio = misses/access 最好是0
    access: The number of times the btree indexes have been accessed.
    hits:The number of times a btree page was in memory
    misses: The number of times a btree page was not in memory
    注意:不要只看一条线
    此处输入图片的描述
    6. Asserts

    意义:The number of regular asserts raised since this process started
    "asserts" : {
    "regular" : ,
    "warning" : ,
    "msg" : ,
    "user" : ,
    "rollovers" :
    },
    不要只看一条线
    此处输入图片的描述
    7. Page faults
    The number of page faults on this process. In non-Windows environments this is hard page faults only.
    越小越好。
    此处输入图片的描述
    ref:link

  • 相关阅读:
    Java中的包装类
    Java中的StringBuffer和StringBuilder
    Java中的String类
    分类模型评价指标:混淆矩阵;ROC曲线;AUC面积
    matlab神经网络多分类(模式识别神经网络nprtool)
    matlabKNN多分类
    matlabSVM多分类:fitcecoc函数
    随机森林和GBDT
    matlab数据预处理
    CDN加速地址URL拿不到,显示“无法访问此网站”
  • 原文地址:https://www.cnblogs.com/no7dw/p/4623325.html
Copyright © 2011-2022 走看看