tsar指标解释
系統(tǒng)模塊
cpu
字段含義
user: 表示CPU執(zhí)行用戶進(jìn)程的時(shí)間,通常期望用戶空間CPU越高越好.
sys: 表示CPU在內(nèi)核運(yùn)行時(shí)間,系統(tǒng)CPU占用率高,表明系統(tǒng)某部分存在瓶頸.通常值越低越好.
wait: CPU在等待I/O操作完成所花費(fèi)的時(shí)間.系統(tǒng)部應(yīng)該花費(fèi)大量時(shí)間來等待I/O操作,否則就說明I/O存在瓶頸.
hirq: 系統(tǒng)處理硬中斷所花費(fèi)的時(shí)間百分比
sirq: 系統(tǒng)處理軟中斷所花費(fèi)的時(shí)間百分比
util: CPU總使用的時(shí)間百分比
nice: 系統(tǒng)調(diào)整進(jìn)程優(yōu)先級所花費(fèi)的時(shí)間百分比
steal: 被強(qiáng)制等待（involuntary wait）虛擬CPU的時(shí)間,此時(shí)hypervisor在為另一個(gè)虛擬處理器服務(wù)
ncpu: CPU的總個(gè)數(shù)
采集方式
CPU的占用率計(jì)算,都是根據(jù)/proc/stat計(jì)數(shù)器文件而來,stat文件的內(nèi)容基本格式是:
cpu 67793686 1353560 66172807 4167536491 2705057 0 195975 609768
cpu0 10529517 944309 11652564 835725059 2150687 0 74605 196726
cpu1 14380773 127146 13908869 832565666 150815 0 31780 108418
cpu是總的信息,cpu0,cpu1等是各個(gè)具體cpu的信息,共有8個(gè)值,單位是ticks,分別是
User time, 67793686
Nice time, 1353560
System time, 66172807
Idle time, 4167536491
Waiting time, 2705057
Hard Irq time, 0
SoftIRQ time, 195975
Steal time, 609768
CPU總時(shí)間=user+system+nice+idle+iowait+irq+softirq+Stl
各個(gè)狀態(tài)的占用=狀態(tài)的cpu時(shí)間％CPU總時(shí)間＊100%
比較特殊的是CPU總使用率的計(jì)算(util),目前的算法是:
util = 1 - idle - iowait - steal
mem
字段含義
free: 空閑的物理內(nèi)存的大小
used: 已經(jīng)使用的內(nèi)存大小
buff: buff使用的內(nèi)存大小,buffer is something that has yet to be "written" to disk.
cach: 操作系統(tǒng)會(huì)把經(jīng)常訪問的東西放在cache中加快執(zhí)行速度,A cache is something that has been "read" from the disk and stored for later use
total: 系統(tǒng)總的內(nèi)存大小
util: 內(nèi)存使用率
采集方法
內(nèi)存的計(jì)數(shù)器在/proc/meminfo,里面有一些關(guān)鍵項(xiàng)
MemTotal: 7680000 kB
MemFree: 815652 kB
Buffers: 1004824 kB
Cached: 4922556 kB
含義就不解釋了,主要介紹一下內(nèi)存使用率的計(jì)算算法:
util = (total - free - buff - cache) / total * 100%
load
字段含義
load1: 一分鐘的系統(tǒng)平均負(fù)載
load5: 五分鐘的系統(tǒng)平均負(fù)載
load15:十五分鐘的系統(tǒng)平均負(fù)載
runq: 在采樣時(shí)刻,運(yùn)行隊(duì)列的任務(wù)的數(shù)目,與/proc/stat的procs_running表示相同意思
plit: 在采樣時(shí)刻,系統(tǒng)中活躍的任務(wù)的個(gè)數(shù)（不包括運(yùn)行已經(jīng)結(jié)束的任務(wù)）
采集方法
/proc/loadavg文件中保存的有負(fù)載相關(guān)的數(shù)據(jù)
0.00 0.01 0.00 1/271 23741
分別是1分鐘負(fù)載,五分鐘負(fù)載,十五分鐘負(fù)載,運(yùn)行進(jìn)程／總進(jìn)程 最大的pid
只需要采集前五個(gè)數(shù)據(jù)既可得到所有信息
注意:只有當(dāng)系統(tǒng)負(fù)載除cpu核數(shù)>1的時(shí)候,系統(tǒng)負(fù)載較高
traffic
字段含義
bytin: 入口流量byte/s
bytout: 出口流量byte/s
pktin: 入口pkt/s
pktout: 出口pkt/s
采集方法
流量的計(jì)數(shù)器信息來自:/proc/net/dev
face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed
lo:1291647853895 811582000 0 0 0 0 0 0 1291647853895 811582000 0 0 0 0 0 0
eth0:853633725380 1122575617 0 0 0 0 0 0 1254282827126 808083790 0 0 0 0 0 0
字段的含義第一行已經(jīng)標(biāo)示出來,每一行代表一個(gè)網(wǎng)卡,tsar主要采集的是出口和入口的bytes／packets
注意tsar只對以eth和em開頭的網(wǎng)卡數(shù)據(jù)進(jìn)行了采集,像lo這種網(wǎng)卡直接就忽略掉了,流量的單位是byte
tcp
字段含義
active:主動(dòng)打開的tcp連接數(shù)目
pasive:被動(dòng)打開的tcp連接數(shù)目
iseg: 收到的tcp報(bào)文數(shù)目
outseg:發(fā)出的tcp報(bào)文數(shù)目
EstRes:Number of resets that have occurred at ESTABLISHED
AtmpFa:Number of failed connection attempts
CurrEs:當(dāng)前狀態(tài)為ESTABLISHED的tcp連接數(shù)
retran:系統(tǒng)的重傳率
采集方法
tcp的相關(guān)計(jì)數(shù)器文件是:/proc/net/snmp
Tcp: RtoAlgorithm RtoMin RtoMax MaxConn ActiveOpens PassiveOpens AttemptFails EstabResets CurrEstab InSegs OutSegs RetransSegs InErrs OutRsts
Tcp: 1 200 120000 -1 31702170 14416937 935062 772446 16 1846056224 1426620266 448823 0 5387732
我們主要關(guān)注其中的ActiveOpens/PassiveOpens/AttemptFails/EstabResets/CurrEstab/InSegs/OutSegs/RetransSegs
主要關(guān)注一下重傳率的計(jì)算方式:
retran = (RetransSegs－last RetransSegs) ／ (OutSegs－last OutSegs) * 100%
udp
字段含義
idgm: 收到的udp報(bào)文數(shù)目
odgm: 發(fā)送的udp報(bào)文數(shù)目
noport:udp協(xié)議層接收到目的地址或目的端口不存在的數(shù)據(jù)包
idmerr:udp層接收到的無效數(shù)據(jù)包的個(gè)數(shù)
采集方法
UDP的數(shù)據(jù)來源文件和TCP一樣,也是在/proc/net/snmp
Udp: InDatagrams NoPorts InErrors OutDatagrams
Udp: 31609577 10708119 0 159885874
io
字段含義
rrqms: The number of read requests merged per second that were issued to the device.
wrqms: The number of write requests merged per second that were issued to the device.
rs: The number of read requests that were issued to the device per second.
ws: The number of write requests that were issued to the device per second.
rsecs: The number of sectors read from the device per second.
wsecs: The number of sectors written to the device per second.
rqsize:The average size (in sectors) of the requests that were issued to the device.
qusize:The average queue length of the requests that were issued to the device.
await: The average time (in milliseconds) for I/O requests issued to the device to be served. This includes the time spent by the requests in queue and the time spent servicing them.
svctm: The average service time (in milliseconds) for I/O requests that were issued to the device.
util: Percentage of CPU time during which I/O requests were issued to the device (bandwidth utilization for the device).Device saturation occurs when this value is close to 100%.
采集方法
IO的計(jì)數(shù)器文件是:/proc/diskstats,比如:
202 0 xvda 12645385 1235409 416827071 59607552 193111576 258112651 3679534806 657719704 0 37341324 717325100
202 1 xvda1 421 2203 3081 9888 155 63 421 1404 0 2608 11292
每一行字段的含義是:
major: 主設(shè)備號
minor: 次設(shè)備號,設(shè)備號是用來區(qū)分磁盤的類型和廠家信息
name: 設(shè)備名稱
rd_ios: 讀完成次數(shù),number of issued reads. This is the total number of reads completed successfully
rd_merges: 合并讀完成次數(shù),為了效率可能會(huì)合并相鄰的讀和寫.從而兩次4K的讀在它最終被處理到磁盤上之前可能會(huì)變成一次8K的讀,才被計(jì)數(shù)（和排隊(duì)）,因此只有一次I/O操作
rd_sectors: 讀扇區(qū)的次數(shù),number of sectors read. This is the total number of sectors read successfully.
rd_ticks: 讀花費(fèi)的毫秒數(shù),number of milliseconds spent reading. This is the total number of milliseconds spent by all reads
wr_ios: 寫完成次數(shù),number of writes completed. This is the total number of writes completed successfully
wr_merges: 合并寫完成次數(shù),number of writes merged Reads and writes which are adjacent to each other may be merged for efficiency. Thus two 4K reads may become one 8K read before it is ultimately handed to the disk, and so it will be counted (and queued) as only one I/O.
wr_sectors: 寫扇區(qū)次數(shù),number of sectors written. This is the total number of sectors written successfully
wr_ticks: 寫花費(fèi)的毫秒數(shù),number of milliseconds spent writing. This is the total number of milliseconds spent by all writes.
cur_ios: 正在處理的輸入/輸出請求數(shù),number of I/Os currently in progress. The only field that should go to zero. Incremented as requests are given to appropriate request_queue_t and decremented as they finish.
ticks: 輸入/輸出操作花費(fèi)的毫秒數(shù)
aveq: 輸入/輸出操作花費(fèi)的加權(quán)毫秒數(shù)
通過這些計(jì)數(shù)器可以算出來上面的每個(gè)字段的值
double n_ios = rd_ios + wr_ios;
double n_ticks = rd_ticks + wr_ticks;
double n_kbytes = (rd_sectors + wr_sectors) / 2;
st_array[0] = rd_merges / (inter * 1.0);
st_array[1] = wr_merges / (inter * 1.0);
st_array[2] = rd_ios / (inter * 1.0);
st_array[3] = wr_ios / (inter * 1.0);
st_array[4] = rd_sectors / (inter * 2.0);
st_array[5] = wr_sectors / (inter * 2.0);
st_array[6] = n_ios ? n_kbytes / n_ios : 0.0;
st_array[7] = aveq / (inter * 1000);
st_array[8] = n_ios ? n_ticks / n_ios : 0.0;
st_array[9] = n_ios ? ticks / n_ios : 0.0;
st_array[10] = ticks / (inter * 10.0); /* percentage! */
/*st_array分別代表tsar顯示的每一個(gè)值*/
注意:
扇區(qū)一般都是512字節(jié),因此有的地方除以2了
ws是指真正落到io設(shè)備上的寫次數(shù), wrqpms是指系統(tǒng)調(diào)用合并的寫次數(shù), 它們之間的大小關(guān)系沒有可比性,因?yàn)椴恢蓝嗌僬埱竽軌虮缓喜?比如發(fā)起了100個(gè)read系統(tǒng)調(diào)用,每個(gè)讀4K,假如這100個(gè)都是連續(xù)的讀,由于硬盤通常允許最大的request為256KB,那么block層會(huì)把這100個(gè)讀請求合并成2個(gè)request,一個(gè)256KB,另一個(gè)144KB,rrqpm/s為100,因?yàn)?00個(gè)request都發(fā)生了合并,不管它最后合并成幾個(gè)；r/s為2,因?yàn)樽詈蟮膔equest數(shù)為2
paritition
字段含義
bfree: 分區(qū)空閑的字節(jié)
bused: 分區(qū)使用中的字節(jié)
btotl: 分區(qū)總的大小
util: 分區(qū)使用率
采集方法
首先通過/etc/mtab獲取到分區(qū)信息,然后通過statfs訪問該分區(qū)的信息,查詢文件系統(tǒng)相關(guān)信息,包含:
struct statfs {
long f_type; /* 文件系統(tǒng)類型 */
long f_bsiz
e; /* 經(jīng)過優(yōu)化的傳輸塊大小 */
long f_blocks; /* 文件系統(tǒng)數(shù)據(jù)塊總數(shù) */
long f_bfree; /* 可用塊數(shù) */
long f_bavail; /* 非超級用戶可獲取的塊數(shù) */
long f_files; /* 文件結(jié)點(diǎn)總數(shù) */
long f_ffree; /* 可用文件結(jié)點(diǎn)數(shù) */
fsid_t f_fsid; /* 文件系統(tǒng)標(biāo)識(shí) */
long f_namelen; /* 文件名的最大長度 */
};
然后就可以計(jì)算出tsar需要的信息,分區(qū)的字節(jié)數(shù)＝塊數(shù)＊塊大小＝f_blocks * f_bsize
pcsw
字段含義
cswch: 進(jìn)程切換次數(shù)
proc: 新建的進(jìn)程數(shù)
采集方法
計(jì)數(shù)器在/proc/stat:
ctxt 19873315174
processes 296444211
分別代表進(jìn)程切換次數(shù),以及進(jìn)程數(shù)
tcpx
字段含義
recvq sendq est twait fwait1 fwait2 lisq lising lisove cnest ndrop edrop rdrop pdrop kdrop
分別代表
tcprecvq tcpsendq tcpest tcptimewait tcpfinwait1 tcpfinwait2 tcplistenq tcplistenincq tcplistenover tcpnconnest tcpnconndrop tcpembdrop tcprexmitdrop tcppersistdrop tcpkadrop
采集方法
計(jì)數(shù)器來自:/proc/net/netstat /proc/net/snmp
里面用到的數(shù)據(jù)有:
TcpExt: SyncookiesSent SyncookiesRecv SyncookiesFailed EmbryonicRsts PruneCalled RcvPruned OfoPruned OutOfWindowIcmps LockDroppedIcmps ArpFilter TW TWRecycled TWKilled PAWSPassive PAWSActive PAWSEstab DelayedACKs DelayedACKLocked DelayedACKLost ListenOverflows ListenDrops TCPPrequeued TCPDirectCopyFromBacklog TCPDirectCopyFromPrequeue TCPPrequeueDropped TCPHPHits TCPHPHitsToUser TCPPureAcks TCPHPAcks TCPRenoRecovery TCPSackRecovery TCPSACKReneging TCPFACKReorder TCPSACKReorder TCPRenoReorder TCPTSReorder TCPFullUndo TCPPartialUndo TCPDSACKUndo TCPLossUndo TCPLoss TCPLostRetransmit TCPRenoFailures TCPSackFailures TCPLossFailures TCPFastRetrans TCPForwardRetrans TCPSlowStartRetrans TCPTimeouts TCPRenoRecoveryFail TCPSackRecoveryFail TCPSchedulerFailed TCPRcvCollapsed TCPDSACKOldSent TCPDSACKOfoSent TCPDSACKRecv TCPDSACKOfoRecv TCPAbortOnSyn TCPAbortOnData TCPAbortOnClose TCPAbortOnMemory TCPAbortOnTimeout TCPAbortOnLinger TCPAbortFailed TCPMemoryPressures
TcpExt: 0 0 0 80 539 0 0 0 0 0 3733709 51268 0 0 0 80 5583301 5966 104803 146887 146887 6500405 39465075 2562794034 0 689613557 2730596 540646233 234702206 0 44187 2066 94 240 0 114 293 1781 7221 60514 185158 2 2 3403 400 107505 5860 24813 174014 0 2966 7 168787 106151 40 32851 2 0 2180 9862 0 15999 0 0 0
具體字段找到并且獲取即可
percpu ncpu
字段含義
字段含義等同cpu模塊,只不過能夠支持采集具體的每一個(gè)cpu的信息
采集方法
等同于cpu模塊
pernic
字段含義
字段含義等同traffic模塊,只不過能夠支持采集具體的每一個(gè)網(wǎng)卡的信息
采集方法
等同于traffic模塊
應(yīng)用模塊
proc
字段含義
user: 某個(gè)進(jìn)程用戶態(tài)cpu消耗
sys: 某個(gè)進(jìn)程系統(tǒng)態(tài)cpu消耗
total:某個(gè)進(jìn)程總的cpu消耗
mem: 某個(gè)進(jìn)程的內(nèi)存消耗百分比
RSS: 某個(gè)進(jìn)程的虛擬內(nèi)存消耗,這是駐留在物理內(nèi)存的一部分.它沒有交換到硬盤.它包括代碼,數(shù)據(jù)和棧
read: 進(jìn)程io讀字節(jié)
write:進(jìn)程的io寫字節(jié)
采集方法
計(jì)數(shù)器文件
/proc/pid/stat:獲取進(jìn)程的cpu信息
/proc/pid/status:獲取進(jìn)程的mem信息
/proc/pid/io:獲取進(jìn)程的讀寫IO信息
注意,需要將采集的進(jìn)程名稱配置在/etc/tsar/tsar.conf總的mod_proc on procname,這樣就會(huì)找到procname的pid,并進(jìn)行數(shù)據(jù)采集
nginx
字段含義
Accept:總共接收的新連接數(shù)目
Handle:總共處理的連接數(shù)目
Reqs:總共產(chǎn)生請求數(shù)目
Active:活躍的連接數(shù),等于read+write+wait
Read:讀取請求數(shù)據(jù)的連接數(shù)目
Write:向用戶寫響應(yīng)數(shù)據(jù)的連接數(shù)目
Wait:長連接等待的連接數(shù)目
Qps:每秒處理的請求數(shù)
Rt:平均響應(yīng)時(shí)間ms
采集方法
通過nginx的采集模塊配置,訪問特定地址,具體參見:https://github.com/taobao/tsar-mod_nginx
location = /nginx_status {
stub_status on;
}
請求到的數(shù)據(jù)是:
Active connections: 1
server accepts handled requests request_time
24 24 7 0
Reading: 0 Writing: 1 Waiting: 0
需要確保nginx配置該location,并且能夠訪問curl http://localhost/nginx_status得到上面的數(shù)據(jù)
如果nginx的端口不是80,則需要在配置文件中指定端口,配置文件是/etc/tsar/tsar.conf,修改mod_nginx on為mod_nginx on 8080
類似的有nginx_code, nginx_domain模塊,相應(yīng)的配置是:
req_status_zone server "$host" 20M;
req_status server;
location /traffic_status {
req_status_show;
}
通過訪問curl http://localhost/traffic_status能夠得到如下字段的數(shù)據(jù)
localhost,0,0,2,2,2,0,0,0,0,0,0,0,0,2,0,0,0,0,0,0,0,0,0,0,0,0,0
請求到的數(shù)據(jù)每個(gè)字段的含義是:
kv 計(jì)算得到的req_status_zone指令定義變量的值,此時(shí)為domain字段
bytes_in_total 從客戶端接收流量總和
bytes_out_total 發(fā)送到客戶端流量總和
conn_total 處理過的連接總數(shù)
req_total 處理過的總請求數(shù)
2xx 2xx請求的總數(shù)
3xx 3xx請求的總數(shù)
4xx 4xx請求的總數(shù)
5xx 5xx請求的總數(shù)
other 其他請求的總數(shù)
rt_total rt的總數(shù)
upstream_req 需要訪問upstream的請求總數(shù)
upstream_rt 訪問upstream的總rt
upstream_tries upstram總訪問次數(shù)
200 200請求的總數(shù)
206 206請求的總數(shù)
302 302請求的總數(shù)
304 304請求的總數(shù)
403 403請求的總數(shù)
404 404請求的總數(shù)
416 416請求的總數(shù)
499 499請求的總數(shù)
500 500請求的總數(shù)
502 502請求的總數(shù)
503 503請求的總數(shù)
504 504請求的總數(shù)
508 508請求的總數(shù)
detail_other 非以上13種status code的請求總數(shù)
如果domain數(shù)量太多,或者端口不是80,需要進(jìn)行專門的配置,配置文件內(nèi)容如下:
port=8080 #指定nginx的端口
top=10 #指定最多采集的域名個(gè)數(shù)，按照請求總個(gè)數(shù)排列
domain=a.com b.com #指定特定需要采集的域名列表,分隔符為空格,逗號,或者制表符
在/etc/tsar/tsar.conf中指定配置文件的路徑:mod_nginx_domain on /tmp/my.conf
nginx_domain_traffic
nginx配置是:
req_status_zone server "$host" 20M;
req_status server;
# req_status_zone_add_indecator 指令：可以在req status輸出的每一行最后添加新的字段
# 這里添加的字段用于統(tǒng)計(jì)nginx的變量: $2xx_bytes_sent, $3xx_bytes_sent, $4xx_bytes_sent, $5xx_bytes_sent
# $2xx_bytes_sent: 請求返回2xx時(shí)候，發(fā)送給客戶端的數(shù)據(jù)量(如果請求非2xx則該變量為0)
req_status_zone_add_indecator server $2xx_bytes_sent $3xx_bytes_sent $4xx_bytes_sent $5xx_bytes_sent;
location /traffic_status {
req_status_show;
}
輸出實(shí)例:
module004033.sqa.cm4 tsar $ tsar --nginx_domain_traffic -li1
Time -----------------localhost:8080----------------- ----------------www.foo.com:8080----------------
Time bytin bytout 2XXout 3XXout 4XXout 5XXout bytin bytout 2XXout 3XXout 4XXout 5XXout
09/01/15-13:45:48 0.00 0.00 0.00 0.00 0.00 0.00 410.1K 16.6M 16.6M 0.00 0.00 0.00
09/01/15-13:45:49 0.00 0.00 0.00 0.00 0.00 0.00 407.8K 16.5M 16.5M 0.00 0.00 0.00
09/01/15-13:45:51 159.0K 287.4K 0.00 0.00 0.00 287.4K 258.6K 10.5M 10.5M 0.00 0.00 0.00
09/01/15-13:45:52 245.5K 443.5K 0.00 0.00 0.00 443.5K 224.2K 9.1M 9.1M 0.00 0.00 0.00
字段含義:
bytin: 收到的請求字節(jié)數(shù)byte/s
bytout: 輸出的應(yīng)答字節(jié)數(shù)byte/s
2XXout: 輸出的2XX應(yīng)答字節(jié)數(shù)byte/s
3XXout: 輸出的3XX應(yīng)答字節(jié)數(shù)byte/s
4XXout: 輸出的4XX應(yīng)答字節(jié)數(shù)byte/s
5XXout: 輸出的5XX應(yīng)答字節(jié)數(shù)byte/s
nginx_ups
用于輸出nginx upstream想關(guān)信息
nginx配置是:
req_status_zone server "$host" 20M;
req_status server;
req_status_zone_add_indecator server $response_fbt_time $upstream_response_fbt_time $upstream_response_length;
location /traffic_status {
req_status_show;
}
輸出實(shí)例:
module004033.sqa.cm4 tsar $ tsar --nginx_ups -li1
Time ----------------------------nginx_ups---------------------------
Time traff qps 4XX 5XX rqps rt fbt ufbt
09/01/15-16:26:29 15.8M 3.9K 3.9K 0.00 0.00 9.7K 9.7K 9.7K
09/01/15-16:26:30 15.8M 3.9K 3.9K 0.00 0.00 9.7K 9.7K 9.7K
09/01/15-16:26:31 4.9M 1.2K 1.2K 0.00 0.00 3.0K 3.0K 3.0K
字段含義:
traff: 后端返回的應(yīng)答body的流量(不包括http應(yīng)答頭部)
qps: 后端qps
rqps: 后端總qps(包含重試的qps + 后端qps)
4XX: 后端返回4XX狀態(tài)碼的qps
5XX: 后端返回5XX狀態(tài)碼的qps
rt: 后端應(yīng)答時(shí)間
fbt: tengine首字節(jié)時(shí)間
ufbt: 后端應(yīng)答首字節(jié)時(shí)間
squid
字段含義
qps: 每秒請求數(shù)
rt: 訪問平均相應(yīng)時(shí)間
r_hit: 請求命中率
b_hit: 字節(jié)命中率
d_hit: 磁盤命中率
m_hit: 內(nèi)存命中率
fdused: Number of file desc currently in use
fdque: Files queued for open
objs: StoreEntries
inmem: StoreEntries with MemObjects
hot: Hot Object Cache Items
size: Mean Object Size
采集方法
訪問squid的mgrinfo信息獲取,有些字段經(jīng)過了一些patch,可能不適用外部版本
haproxy
字段含義
stat: 狀態(tài),1正常
uptime:啟動(dòng)持續(xù)時(shí)間
conns: 總的連接數(shù)
qps: 每秒請求數(shù)
hit: haproxy開啟cache時(shí)的命中率
rt: 平均響應(yīng)時(shí)間ms
采集方法
haproxy經(jīng)過了patch,能夠在多進(jìn)程模式下進(jìn)行統(tǒng)計(jì)信息的匯總,然后通過haproxy的本地訪問其狀態(tài)頁面admin分析得到
lvs
字段含義
stat: lvs狀態(tài),1正常
conns: 總的連接數(shù)
pktin: 收到的包數(shù)
pktout:發(fā)出的包數(shù)
bytin: 收到的字節(jié)數(shù)
bytout:發(fā)出的字節(jié)數(shù)
采集方法
訪問lvs的統(tǒng)計(jì)文件:/proc/net/ip_vs_stats
apache
參見:https://github.com/kongjian/tsar-apache
tcprt
私有應(yīng)用,略
swift
私有應(yīng)用,略
cgcpu/cgmem/cgblkio
私有應(yīng)用,略
trafficserver
待補(bǔ)充
tmd
私有應(yīng)用,略
總結(jié)
- 上一篇： java中组合_java中组合模式详解和
- 下一篇： MySQL 使用AVG聚合函数时，保留两