好了，今天接着來學習性能，在今天開始前，我今天在網上又看到了理發師經典模型，這里稍微提一下，詳情可以百度哈，下面這張圖是網上找到的經典場景性能相關的圖，大致說明下：
這張圖中展示的是1個標准的軟件性能模型。在圖中有三條曲線，分別表示資源的利用情況（Utilization，包括硬件資源和軟件資源）、吞吐量（Throughput，這里是指每秒事務數）以及響應時間（Response Time）。圖中坐標軸的橫軸從左到右表現了並發用戶數（Number of Concurrent Users）的不斷增長。
在這張圖中我們可以看到，最開始，隨着並發用戶數的增長，資源占用率和吞吐量會相應的增長，但是響應時間的變化不大；不過當並發用戶數增長到一定程度后，資源占用達到飽和，吞吐量增長明顯放緩甚至停止增長，而響應時間卻進一步延長。如果並發用戶數繼續增長，你會發現軟硬件資源占用繼續維持在飽和狀態，但是吞吐量開始下降，響應時間明顯的超出了用戶可接受的范圍，並且最終導致用戶放棄了這次請求甚至離開。
根據這種性能表現，圖中划分了三個區域，分別是Light Load（較輕的壓力）、Heavy Load（較重的壓力）和Buckle Zone（用戶無法忍受並放棄請求）。在Light Load和Heavy Load 兩個區域交界處的並發用戶數，我們稱為“最佳並發用戶數（The Optimum Number of Concurrent Users）”，而Heavy Load和Buckle Zone兩個區域交界處的並發用戶數則稱為“最大並發用戶數（The Maximum Number of Concurrent Users）”。
我們在進行性能測試時，首先需要確定的就是性能指標，然后開始性能建模，接着確定性能方案，最后進行性能場景。
下面是性能測試的大致過程：
下面是性能建模的各個接口單個小時的指標計算，最后發現單個線程就能滿足TPS了。（所以說用戶不等於線程數）
上面稍微提了一下性能的過程和指標計算，這些會在后面詳細講解，今天的主要目的實際上是來做性能監控的，性能監控的數據我們用influxdb進行保存，然后數據展示用grafana。
- 這里我們先來安裝下influxdb，influxdb 1.5.4 下載地址 ：https://www.influxdata.com/ ， 下載之后通過yum localinstall安裝rpm包（跟着官網干的）
下圖是啟動influxdb然后通過influx命令進入influxdb數據庫，創建jmeter數據庫，show MEASURENENTS （這個是查看所有鍵）。在頁面上可以試着訪問下ip+8086，如果出現404 page not found證明OK，因為是通用http請求的。
- 創建一個新的管理員用戶
CREATE USER <username> WITH PASSWORD '<password>' WITH ALL PRIVILEGES
- 為一個已有用戶授權管理員權限
GRANT ALL PRIVILEGES TO <username>
2.安裝grafana，下載地址：https://grafana.com/grafana/download，下載之后通過yum localinstall安裝rpm包。
安裝完成后啟動服務：systemctl start grafana-server，記住要關閉防火牆喔，systemctl stop firewalld（臨時關閉）和systemctl disable firewalld（禁止開機啟動）
然后訪問，ip+3000端口，默認用戶名和密碼是admin;
安裝好了grafana之后，我們可以配置一下grafana從influxdb中讀取數據，從設置->Data Sources然后進行如下的配置：
3.配置完后，我們可以通過jmeter通過Backend Listener將測試的數據寫入influxdb的jmeter表中，如下圖：這里用三個java請求跑一下看下效果
我們可以查詢看下jmeter表里面有哪些字段及意思：
time 時間戳 application jmeter中指定的application的名稱 avg 平均響應時間 count 所有產生東西的統計 countError 錯誤統計 endedT 結束的線程 hit max 最大的響應時間 maxAT 最大的活動線程數 meanAT min 最小的響應時間 minAT pct90.0 90%的響應時間 pct95.0 95%的響應時間 pct99.0 99%的響應時間 startedT 開始的線程 statut 狀態 transaction 事務
4.我們通過New dashboard -> add -> Singlestat來添加剛剛的請求數，剛剛15個線程跑三個java請求各100次，應該是4500次，但是因為兩個請求又包成了一層事務，所以總的請求次數為6000：
下圖是三個請求的TPS：
最終模仿着別人的寫出來了自己的模板，效果如下：
如果需要源文件可以私撩我 ，對於添加各種數據圖表的話，大家可以查看該地址：https://testerhome.com/topics/11256
5.上面是從jmeter結果中讀取的數據，主要還是請求性能的參數指標，那么我們在真正的性能測試過程中，我們經常還需要統計服務器的資源使用情況，那么這個時候我們怎么將服務器的資源使用情況也寫進來呢？
我們可以查看influxdb官網的API接口，通過curl -i -XPOST http://localhost:8086/query --data-urlencode "q=CREATE DATABASE mydb"可以創建mydb數據庫
然后我們可以先將監控數據保存在txt文檔內，最后再將txt監控數據寫入到influxdb數據庫中，最終通過grafana展示出來。
寫入influxdb數據庫的命令：
curl -i -XPOST 'http://localhost:8086/write?db=mydb' --data-binary 'cpu_load_short,host=server01,region=us-west value=0.64 1434055562000000000'- 或：curl -i -XPOST 'http://localhost:8086/write?db=mydb' --data-binary @cpu_data.txt
對於linux的指標監控我們用的是vmstat，這里來簡單說明一下各個字段的意思：
字段說明：
Procs（進程）
- r：運行隊列中進程數量，以8核為例，如果隊列數量超過8，證明有排隊的，但這不代表壓力很大，因為很快就能處理完了，如果該參數長期大於和等於邏輯cpu個數，則CPU資源可能存在較大的瓶頸。
- b： 等待IO的進程數量。意味着進程被阻塞。主要是指被資源阻塞的進程對列數（比如IO資源、頁面調度等），當這個值較大時，需要根據應用程序來進行分析
Memory（內存）
- swpd: 使用虛擬內存大小，如果swpd的值不為0，但是SI，SO的值長期為0，這種情況不會影響系統性能。
- free: 空閑物理內存大小。
- buff: 用作緩沖的內存大小。
- cache: 用作緩存的內存大小，如果cache的值大的時候，說明cache處的文件數多，如果頻繁訪問到的文件都能被cache處，那么磁盤的讀IO bi會非常小。
Swap
- si: 每秒從交換區寫到內存的大小，由磁盤調入內存。
- so: 每秒寫入交換區的內存大小，由內存調入磁盤。
注意：內存夠用的時候，這2個值都是0，如果這2個值長期大於0時，系統性能會受到影響，磁盤IO和CPU資源都會被消耗。有些朋友看到空閑內存（free）很少的或接近於0時，就認為內存不夠用了，不能光看這一點，還要結合si和so，如果free很少，但是si和so也很少（大多時候是0），那么不用擔心，系統性能這時不會受到影響的。
IO（現在的Linux版本塊的大小為1kb）
- bi: 每秒讀取的塊數
- bo: 每秒寫入的塊數
注意：隨機磁盤讀寫的時候，這2個值越大（如超出1024k)，能看到CPU在IO等待的值也會越大。
system（系統）
- in: 每秒中斷數，包括時鍾中斷。（分為軟中斷（訪問應用軟件超時等）和硬中斷（訪問硬件報錯，硬中斷+1））
- cs: 每秒上下文切換數。（進程和線程的交互，經驗來講，2C4G最好小於2W次，4C8G在6W~8W，8核16G是在20W到30W）
注意：上面2個值越大，會看到由內核消耗的CPU時間會越大。
CPU（以百分比表示）
- us: 用戶進程執行時間百分比(user time)
us的值比較高時，說明用戶進程消耗的CPU時間多，但是如果長期超50%的使用，那么我們就該考慮優化程序算法或者進行加速。
- sy: 內核系統進程執行時間百分比(system time)
sy的值高時，說明系統內核消耗的CPU資源多，這並不是良性表現，我們應該檢查原因。
- id: 空閑時間百分比
- wa: IO等待時間百分比
wa的值高時，說明IO等待比較嚴重，這可能由於磁盤大量作隨機訪問造成，也有可能磁盤出現瓶頸（塊操作）。
- st: 當虛擬機1資源不夠的時候會去偷宿主機資源（宿主機資源不夠情況可以再去虛擬機2拿），最終給虛擬機1，這個過程就是st
6.字段說明完了之后，我們就來寫shell腳本通過vmstat去收集數據然后再插入influxdb數據庫中，最后再從grafana中通過sql語句讀取出數據展現在圖表中。
首先，我先寫個sh文件，通過vmstat收集數據然后post到influxdb中，sh腳本內容如下：
#!/bin/bash while true;do value=`vmstat 1 2|tail -n 1` vm_r=`echo $value|cut -d " " -f1` vm_b=`echo $value|cut -d " " -f2` vm_swpd=`echo $value|cut -d " " -f3` vm_free=`echo $value|cut -d " " -f4` vm_buff=`echo $value|cut -d " " -f5` vm_cache=`echo $value|cut -d " " -f6` vm_si=`echo $value|cut -d " " -f7` vm_so=`echo $value|cut -d " " -f8` vm_bi=`echo $value|cut -d " " -f9` vm_bo=`echo $value|cut -d " " -f10` vm_in=`echo $value|cut -d " " -f11` vm_cs=`echo $value|cut -d " " -f12` vm_us=`echo $value|cut -d " " -f13` vm_sy=`echo $value|cut -d " " -f14` vm_id=`echo $value|cut -d " " -f15` vm_wa=`echo $value|cut -d " " -f16` vm_st=`echo $value|cut -d " " -f17` ip=`ip addr |grep eno16777736|tail -n 1|awk '{print $2}'|awk -F '/' '{ print $1}'` curl -i -XPOST "http://10.10.18.131:8086/write?db=vmstat" --data-binary "linux_load_count,ip=$ip r=$vm_r linux_load_count,ip=$ip b=$vm_b linux_load_count,ip=$ip swpd=$vm_swpd linux_load_count,ip=$ip free=$vm_free linux_load_count,ip=$ip buff=$vm_buff linux_load_count,ip=$ip cache=$vm_cache linux_load_count,ip=$ip si=$vm_si linux_load_count,ip=$ip so=$vm_so linux_load_count,ip=$ip bi=$vm_bi linux_load_count,ip=$ip bo=$vm_bo linux_load_count,ip=$ip in=$vm_in linux_load_count,ip=$ip cs=$vm_cs linux_load_count,ip=$ip us=$vm_us linux_load_count,ip=$ip sy=$vm_sy linux_load_count,ip=$ip id=$vm_id linux_load_count,ip=$ip wa=$vm_wa linux_load_count,ip=$ip st=$vm_st" sleep 3 done
推送完成后，我們就去grafana寫sql展示（記得新建一個data source），效果圖如下：
其實不是很難，我把這個模板的json貼出來，有興趣的可以一起深入研究研究。
下面是linux監控的json模板：
{
"__inputs": [
{
"name": "DS_VMSTAT",
"label": "vmstat",
"description": "",
"type": "datasource",
"pluginId": "influxdb",
"pluginName": "InfluxDB"
}
],
"__requires": [
{
"type": "grafana",
"id": "grafana",
"name": "Grafana",
"version": "5.2.1"
},
{
"type": "panel",
"id": "graph",
"name": "Graph",
"version": "5.0.0"
},
{
"type": "datasource",
"id": "influxdb",
"name": "InfluxDB",
"version": "5.0.0"
}
],
"annotations": {
"list": [
{
"builtIn": 1,
"datasource": "-- Grafana --",
"enable": true,
"hide": true,
"iconColor": "rgba(0, 211, 255, 1)",
"name": "Annotations & Alerts",
"type": "dashboard"
}
]
},
"description": "監控linux機器的資源使用情況",
"editable": true,
"gnetId": null,
"graphTooltip": 0,
"id": null,
"links": [],
"panels": [
{
"aliasColors": {},
"bars": false,
"dashLength": 10,
"dashes": false,
"datasource": "${DS_VMSTAT}",
"description": "us: 用戶進程執行時間百分比(user time)\nus的值比較高時，說明用戶進程消耗的CPU時間多，但是如果長期超50%的使用，那么我們就該考慮優化程序算法或者進行加速。\n\nsy: 內核系統進程執行時間百分比(system time)\nsy的值高時，說明系統內核消耗的CPU資源多，這並不是良性表現，我們應該檢查原因。\n\nid: 空閑時間百分比\nwa: IO等待時間百分比\nwa的值高時，說明IO等待比較嚴重，這可能由於磁盤大量作隨機訪問造成，也有可能磁盤出現瓶頸（塊操作）。\n\nst: 當虛擬機1資源不夠的時候會去偷宿主機資源（宿主機資源不夠情況可以再去虛擬機2拿），最終給虛擬機1，這個過程就是st",
"fill": 2,
"gridPos": {
"h": 9,
"w": 12,
"x": 0,
"y": 0
},
"id": 2,
"legend": {
"avg": false,
"current": false,
"max": false,
"min": false,
"show": true,
"total": false,
"values": false
},
"lines": true,
"linewidth": 2,
"links": [],
"nullPointMode": "null",
"percentage": false,
"pointradius": 5,
"points": false,
"renderer": "flot",
"seriesOverrides": [],
"spaceLength": 10,
"stack": false,
"steppedLine": false,
"targets": [
{
"alias": "$col",
"groupBy": [
{
"params": [
"$__interval"
],
"type": "time"
},
{
"params": [
"null"
],
"type": "fill"
}
],
"orderByTime": "ASC",
"policy": "default",
"query": "SELECT us from linux_load_count WHERE $timeFilter fill(null)",
"rawQuery": true,
"refId": "A",
"resultFormat": "time_series",
"select": [
[
{
"params": [
"value"
],
"type": "field"
},
{
"params": [],
"type": "mean"
}
]
],
"tags": []
},
{
"alias": "$col",
"groupBy": [
{
"params": [
"$__interval"
],
"type": "time"
},
{
"params": [
"null"
],
"type": "fill"
}
],
"orderByTime": "ASC",
"policy": "default",
"query": "SELECT sy from linux_load_count WHERE $timeFilter fill(null)",
"rawQuery": true,
"refId": "B",
"resultFormat": "time_series",
"select": [
[
{
"params": [
"value"
],
"type": "field"
},
{
"params": [],
"type": "mean"
}
]
],
"tags": []
},
{
"alias": "$col",
"groupBy": [
{
"params": [
"$__interval"
],
"type": "time"
},
{
"params": [
"null"
],
"type": "fill"
}
],
"orderByTime": "ASC",
"policy": "default",
"query": "SELECT id from linux_load_count WHERE $timeFilter fill(null)",
"rawQuery": true,
"refId": "C",
"resultFormat": "time_series",
"select": [
[
{
"params": [
"value"
],
"type": "field"
},
{
"params": [],
"type": "mean"
}
]
],
"tags": []
},
{
"alias": "$col",
"groupBy": [
{
"params": [
"$__interval"
],
"type": "time"
},
{
"params": [
"null"
],
"type": "fill"
}
],
"orderByTime": "ASC",
"policy": "default",
"query": "SELECT wa from linux_load_count WHERE $timeFilter fill(null)",
"rawQuery": true,
"refId": "D",
"resultFormat": "time_series",
"select": [
[
{
"params": [
"value"
],
"type": "field"
},
{
"params": [],
"type": "mean"
}
]
],
"tags": []
},
{
"alias": "$col",
"groupBy": [
{
"params": [
"$__interval"
],
"type": "time"
},
{
"params": [
"null"
],
"type": "fill"
}
],
"orderByTime": "ASC",
"policy": "default",
"query": "SELECT st from linux_load_count WHERE $timeFilter fill(null)",
"rawQuery": true,
"refId": "E",
"resultFormat": "time_series",
"select": [
[
{
"params": [
"value"
],
"type": "field"
},
{
"params": [],
"type": "mean"
}
]
],
"tags": []
}
],
"thresholds": [],
"timeFrom": null,
"timeShift": null,
"title": "CPU Load",
"tooltip": {
"shared": true,
"sort": 0,
"value_type": "individual"
},
"type": "graph",
"xaxis": {
"buckets": null,
"mode": "time",
"name": null,
"show": true,
"values": []
},
"yaxes": [
{
"format": "percent",
"label": null,
"logBase": 1,
"max": null,
"min": null,
"show": true
},
{
"format": "short",
"label": null,
"logBase": 1,
"max": null,
"min": null,
"show": true
}
],
"yaxis": {
"align": false,
"alignLevel": null
}
},
{
"aliasColors": {},
"bars": false,
"dashLength": 10,
"dashes": false,
"datasource": "${DS_VMSTAT}",
"description": "swpd: 使用虛擬內存大小，如果swpd的值不為0，但是SI，SO的值長期為0，這種情況不會影響系統性能。\n\nfree: 空閑物理內存大小。\n\nbuff: 用作緩沖的內存大小。\n\ncache: 用作緩存的內存大小，如果cache的值大的時候，說明cache處的文件數多，如果頻繁訪問到的文件都能被cache處，那么磁盤的讀IO bi會非常小。\n\nsi: 每秒從交換區寫到內存的大小，由磁盤調入內存。\n\nso: 每秒寫入交換區的內存大小，由內存調入磁盤.\n\n注意：內存夠用的時候，這2個值都是0，如果這2個值長期大於0時，系統性能會受到影響，磁盤IO和CPU資源都會被消耗。有些朋友看到空閑內存（free）很少的或接近於0時，就認為內存不夠用了，不能光看這一點，還要結合si和so，如果free很少，但是si和so也很少（大多時候是0），那么不用擔心，系統性能這時不會受到影響的。",
"fill": 2,
"gridPos": {
"h": 9,
"w": 12,
"x": 12,
"y": 0
},
"id": 4,
"legend": {
"avg": false,
"current": false,
"max": false,
"min": false,
"show": true,
"total": false,
"values": false
},
"lines": true,
"linewidth": 2,
"links": [],
"nullPointMode": "null",
"percentage": false,
"pointradius": 5,
"points": false,
"renderer": "flot",
"seriesOverrides": [],
"spaceLength": 10,
"stack": false,
"steppedLine": false,
"targets": [
{
"alias": "memory_$col",
"groupBy": [
{
"params": [
"$__interval"
],
"type": "time"
},
{
"params": [
"null"
],
"type": "fill"
}
],
"orderByTime": "ASC",
"policy": "default",
"query": "SELECT free from linux_load_count WHERE $timeFilter fill(null)",
"rawQuery": true,
"refId": "A",
"resultFormat": "time_series",
"select": [
[
{
"params": [
"value"
],
"type": "field"
},
{
"params": [],
"type": "mean"
}
]
],
"tags": []
},
{
"alias": "memory_$col",
"groupBy": [
{
"params": [
"$__interval"
],
"type": "time"
},
{
"params": [
"null"
],
"type": "fill"
}
],
"orderByTime": "ASC",
"policy": "default",
"query": "SELECT buff from linux_load_count WHERE $timeFilter fill(null)",
"rawQuery": true,
"refId": "B",
"resultFormat": "time_series",
"select": [
[
{
"params": [
"value"
],
"type": "field"
},
{
"params": [],
"type": "mean"
}
]
],
"tags": []
},
{
"alias": "memory_$col",
"groupBy": [
{
"params": [
"$__interval"
],
"type": "time"
},
{
"params": [
"null"
],
"type": "fill"
}
],
"orderByTime": "ASC",
"policy": "default",
"query": "SELECT cache from linux_load_count WHERE $timeFilter fill(null)",
"rawQuery": true,
"refId": "C",
"resultFormat": "time_series",
"select": [
[
{
"params": [
"value"
],
"type": "field"
},
{
"params": [],
"type": "mean"
}
]
],
"tags": []
},
{
"alias": "memory_$col",
"groupBy": [
{
"params": [
"$__interval"
],
"type": "time"
},
{
"params": [
"null"
],
"type": "fill"
}
],
"orderByTime": "ASC",
"policy": "default",
"query": "SELECT swpd from linux_load_count WHERE $timeFilter fill(null)",
"rawQuery": true,
"refId": "D",
"resultFormat": "time_series",
"select": [
[
{
"params": [
"value"
],
"type": "field"
},
{
"params": [],
"type": "mean"
}
]
],
"tags": []
},
{
"alias": "memory_$col",
"groupBy": [
{
"params": [
"$__interval"
],
"type": "time"
},
{
"params": [
"null"
],
"type": "fill"
}
],
"orderByTime": "ASC",
"policy": "default",
"query": "SELECT si from linux_load_count WHERE $timeFilter fill(null)",
"rawQuery": true,
"refId": "E",
"resultFormat": "time_series",
"select": [
[
{
"params": [
"value"
],
"type": "field"
},
{
"params": [],
"type": "mean"
}
]
],
"tags": []
},
{
"alias": "memory_$col",
"groupBy": [
{
"params": [
"$__interval"
],
"type": "time"
},
{
"params": [
"null"
],
"type": "fill"
}
],
"orderByTime": "ASC",
"policy": "default",
"query": "SELECT so from linux_load_count WHERE $timeFilter fill(null)",
"rawQuery": true,
"refId": "F",
"resultFormat": "time_series",
"select": [
[
{
"params": [
"value"
],
"type": "field"
},
{
"params": [],
"type": "mean"
}
]
],
"tags": []
}
],
"thresholds": [],
"timeFrom": null,
"timeShift": null,
"title": "Memory Load",
"tooltip": {
"shared": true,
"sort": 0,
"value_type": "individual"
},
"type": "graph",
"xaxis": {
"buckets": null,
"mode": "time",
"name": null,
"show": true,
"values": []
},
"yaxes": [
{
"format": "short",
"label": null,
"logBase": 1,
"max": null,
"min": null,
"show": true
},
{
"format": "short",
"label": null,
"logBase": 1,
"max": null,
"min": null,
"show": true
}
],
"yaxis": {
"align": false,
"alignLevel": null
}
},
{
"aliasColors": {},
"bars": false,
"dashLength": 10,
"dashes": false,
"datasource": "${DS_VMSTAT}",
"description": "r：運行隊列中進程數量，以8核為例，如果隊列數量超過8，證明有排隊的，但這不代表壓力很大，因為很快就能處理完了，如果該參數長期大於和等於邏輯cpu個數，則CPU資源可能存在較大的瓶頸。\n\nb： 等待IO的進程數量。意味着進程被阻塞。主要是指被資源阻塞的進程對列數（比如IO資源、頁面調度等），當這個值較大時，需要根據應用程序來進行分析",
"fill": 2,
"gridPos": {
"h": 9,
"w": 8,
"x": 0,
"y": 9
},
"id": 10,
"legend": {
"avg": false,
"current": false,
"max": false,
"min": false,
"show": true,
"total": false,
"values": false
},
"lines": true,
"linewidth": 2,
"links": [],
"nullPointMode": "null",
"percentage": false,
"pointradius": 5,
"points": false,
"renderer": "flot",
"seriesOverrides": [],
"spaceLength": 10,
"stack": false,
"steppedLine": false,
"targets": [
{
"alias": "$col",
"groupBy": [
{
"params": [
"$__interval"
],
"type": "time"
},
{
"params": [
"null"
],
"type": "fill"
}
],
"orderByTime": "ASC",
"policy": "default",
"query": "select \"r\" as process_queue from linux_load_count WHERE $timeFilter fill(null)",
"rawQuery": true,
"refId": "A",
"resultFormat": "time_series",
"select": [
[
{
"params": [
"value"
],
"type": "field"
},
{
"params": [],
"type": "mean"
}
]
],
"tags": []
},
{
"alias": "$col",
"groupBy": [
{
"params": [
"$__interval"
],
"type": "time"
},
{
"params": [
"null"
],
"type": "fill"
}
],
"orderByTime": "ASC",
"policy": "default",
"query": "select \"b\" as block_queue from linux_load_count WHERE $timeFilter fill(null)",
"rawQuery": true,
"refId": "B",
"resultFormat": "time_series",
"select": [
[
{
"params": [
"value"
],
"type": "field"
},
{
"params": [],
"type": "mean"
}
]
],
"tags": []
}
],
"thresholds": [],
"timeFrom": null,
"timeShift": null,
"title": "Procs_queue",
"tooltip": {
"shared": true,
"sort": 0,
"value_type": "individual"
},
"type": "graph",
"xaxis": {
"buckets": null,
"mode": "time",
"name": null,
"show": true,
"values": []
},
"yaxes": [
{
"format": "short",
"label": null,
"logBase": 1,
"max": null,
"min": null,
"show": true
},
{
"format": "short",
"label": null,
"logBase": 1,
"max": null,
"min": null,
"show": true
}
],
"yaxis": {
"align": false,
"alignLevel": null
}
},
{
"aliasColors": {},
"bars": false,
"dashLength": 10,
"dashes": false,
"datasource": "${DS_VMSTAT}",
"description": "in: 每秒中斷數，包括時鍾中斷。（分為軟中斷（訪問應用軟件超時等）和硬中斷（訪問硬件報錯，硬中斷+1））\n\ncs: 每秒上下文切換數。（進程和線程的交互，經驗來講，2C4G最好小於2W次，4C8G在6W~8W，8核16G是在20W到30W）\n\n注意：上面2個值越大，會看到由內核消耗的CPU時間會越大。",
"fill": 2,
"gridPos": {
"h": 9,
"w": 8,
"x": 8,
"y": 9
},
"id": 8,
"legend": {
"avg": false,
"current": false,
"max": false,
"min": false,
"show": true,
"total": false,
"values": false
},
"lines": true,
"linewidth": 2,
"links": [],
"nullPointMode": "null",
"percentage": false,
"pointradius": 5,
"points": false,
"renderer": "flot",
"seriesOverrides": [],
"spaceLength": 10,
"stack": false,
"steppedLine": false,
"targets": [
{
"alias": "$col",
"groupBy": [
{
"params": [
"$__interval"
],
"type": "time"
},
{
"params": [
"null"
],
"type": "fill"
}
],
"orderByTime": "ASC",
"policy": "default",
"query": "select \"in\" as Interruption from linux_load_count WHERE $timeFilter fill(null)",
"rawQuery": true,
"refId": "A",
"resultFormat": "time_series",
"select": [
[
{
"params": [
"value"
],
"type": "field"
},
{
"params": [],
"type": "mean"
}
]
],
"tags": []
},
{
"alias": "$col",
"groupBy": [
{
"params": [
"$__interval"
],
"type": "time"
},
{
"params": [
"null"
],
"type": "fill"
}
],
"orderByTime": "ASC",
"policy": "default",
"query": "select \"cs\" as context_switch from linux_load_count WHERE $timeFilter fill(null)",
"rawQuery": true,
"refId": "B",
"resultFormat": "time_series",
"select": [
[
{
"params": [
"value"
],
"type": "field"
},
{
"params": [],
"type": "mean"
}
]
],
"tags": []
}
],
"thresholds": [],
"timeFrom": null,
"timeShift": null,
"title": "System",
"tooltip": {
"shared": true,
"sort": 0,
"value_type": "individual"
},
"type": "graph",
"xaxis": {
"buckets": null,
"mode": "time",
"name": null,
"show": true,
"values": []
},
"yaxes": [
{
"format": "short",
"label": null,
"logBase": 1,
"max": null,
"min": null,
"show": true
},
{
"format": "short",
"label": null,
"logBase": 1,
"max": null,
"min": null,
"show": true
}
],
"yaxis": {
"align": false,
"alignLevel": null
}
},
{
"aliasColors": {},
"bars": false,
"dashLength": 10,
"dashes": false,
"datasource": "${DS_VMSTAT}",
"description": "bi: 每秒讀取的塊數\n\nbo: 每秒寫入的塊數\n\n注意：隨機磁盤讀寫的時候，這2個值越大（如超出1024k)，能看到CPU在IO等待的值也會越大。",
"fill": 2,
"gridPos": {
"h": 9,
"w": 8,
"x": 16,
"y": 9
},
"id": 6,
"legend": {
"avg": false,
"current": false,
"max": false,
"min": false,
"show": true,
"total": false,
"values": false
},
"lines": true,
"linewidth": 2,
"links": [],
"nullPointMode": "null",
"percentage": false,
"pointradius": 5,
"points": false,
"renderer": "flot",
"seriesOverrides": [],
"spaceLength": 10,
"stack": false,
"steppedLine": false,
"targets": [
{
"alias": "$col",
"groupBy": [
{
"params": [
"$__interval"
],
"type": "time"
},
{
"params": [
"null"
],
"type": "fill"
}
],
"orderByTime": "ASC",
"policy": "default",
"query": "select bi as block_in from linux_load_count WHERE $timeFilter fill(null)",
"rawQuery": true,
"refId": "A",
"resultFormat": "time_series",
"select": [
[
{
"params": [
"value"
],
"type": "field"
},
{
"params": [],
"type": "mean"
}
]
],
"tags": []
},
{
"alias": "$col",
"groupBy": [
{
"params": [
"$__interval"
],
"type": "time"
},
{
"params": [
"null"
],
"type": "fill"
}
],
"orderByTime": "ASC",
"policy": "default",
"query": "select bo as block_out from linux_load_count WHERE $timeFilter fill(null)",
"rawQuery": true,
"refId": "B",
"resultFormat": "time_series",
"select": [
[
{
"params": [
"value"
],
"type": "field"
},
{
"params": [],
"type": "mean"
}
]
],
"tags": []
}
],
"thresholds": [],
"timeFrom": null,
"timeShift": null,
"title": "IO_Load",
"tooltip": {
"shared": true,
"sort": 0,
"value_type": "individual"
},
"type": "graph",
"xaxis": {
"buckets": null,
"mode": "time",
"name": null,
"show": true,
"values": []
},
"yaxes": [
{
"format": "kbytes",
"label": null,
"logBase": 1,
"max": null,
"min": null,
"show": true
},
{
"format": "short",
"label": null,
"logBase": 1,
"max": null,
"min": null,
"show": true
}
],
"yaxis": {
"align": false,
"alignLevel": null
}
}
],
"refresh": false,
"schemaVersion": 16,
"style": "dark",
"tags": [],
"templating": {
"list": []
},
"time": {
"from": "2018-07-25T02:07:25.151Z",
"to": "2018-07-25T02:17:25.151Z"
},
"timepicker": {
"refresh_intervals": [
"5s",
"10s",
"30s",
"1m",
"5m",
"15m",
"30m",
"1h",
"2h",
"1d"
],
"time_options": [
"5m",
"15m",
"1h",
"6h",
"12h",
"24h",
"2d",
"7d",
"30d"
]
},
"timezone": "",
"title": "linux_load",
"uid": "zeTsFdKik",
"version": 10
}