Loadrunner 结果分析向导1.docx
- 文档编号:11345849
- 上传时间:2023-02-28
- 格式:DOCX
- 页数:16
- 大小:328.25KB
Loadrunner 结果分析向导1.docx
《Loadrunner 结果分析向导1.docx》由会员分享,可在线阅读,更多相关《Loadrunner 结果分析向导1.docx(16页珍藏版)》请在冰豆网上搜索。
Loadrunner结果分析向导1
LoadrunnerResultsAnalysis:
Contents
AnalysisSummaryPage3
RunningVusers5
RunningVusers5
HitsperSecond6
Throughput7
HitsperSecond-Throughput8
AverageTransactionResponseTime9
TransactionsperSecond(Passed/Failed)10
TotalTransactionsperSecond11
TransactionSummaryGraph12
TransactionPerformanceSummary13
TransactionResponseTime(UnderLoad)14
TransactionResponseTime(UnderLoad)14
TransactionResponseTime(Percentile)15
HTTPStatusCodeSummary16
HTTPResponsesperSecond17
RetriesperSecond18
Connections19
ConnectionsperSecond20
ErrorStatistics21
ErrorStatistics(byDescription)22
ErrorsperSecond23
ErrorsperSecond(byDescription)24
ErrorsperSecond(byDescription)24
ErrorsperSecond–RunningVusers25
Glossary26
AnalysisSummaryPage
TheAnalysisSummarypageisdividedintothreemainsections,thestatisticssummary,transactionsummary,andHTTPresponsessummary.Thestatisticssummarylistsoverallstatisticsofinterestinthescenario.
YoucancomparemaximumrunningVusersandthetotaldurationofthescenario(herethedurationis17:
08:
38–17:
54:
35,or46minutes)againstthroughputandhitsstatistics.BasedonVusersandduration,wastheperformanceoftheserver/applicationacceptablegiventhethroughputandhits/secondstatistics?
Youwillhavetohavesomeknowledgeofyourserver/databasecapabilitiesandconfigurationtomakethisdecision.
ThenextsectionoftheAnalysisSummaryistheTransactionSummary,containingstatisticsonindividualtransactions,theirresponsetimes(minimum,average,maximum,andstandarddeviation).
The“90Percent”statisticindicatesthemaximumresponsetimeofninetypercentforthetransactionsinthatrow.Basedonthestatisticsabove,youcouldsaythat90%ofthe“A_MainPage_AFI…”transactionshadamaximumresponsetimeof94.845seconds.
Passedandfailedtransactionsarealsolistedinthissection.Transactionswiththe“Z_”prefixcontainthesumofallstatisticsfortransactionsusedinthatscript.Thetotalpassedtransactionsinthe“Z_”transactionwillbeatmostequaltothesingletransactionwiththefewestpasses,excludingvuser_init/endtransactions.Thisisbecausethe“Z_”transactionrecordsstatisticsfortheentirescriptflow,notjustasingletransaction.IfaVusersuccessfullyexecutesallthestepsinascript,fromstarttofinish,onlythenwillthenumberofpassesforthe“Z_”transactionincrease.However,iffailuresoccuratanypointinthescript,thenumberoffailuresforthetransactioninwhichthefailureoccurredwillincrease,andthenumberoffailuresforthe“Z_”transactionwillincreaseaswell.
Inthestatisticsabove,thetotalnumberofpassesinthe“Z_”transactionwas776,correspondingtothesingletransactionwiththefewestpasses,“E_ResultsDisplay…”,whereasthetotalnumberoffailuresfor“Z_”isthesumoffailuresfortheentiregroup.
ThefinalsectionintheAnalysisSummaryistheHTTPResponsesSummary,whichcontainsthetotalandpersecondstatisticsforallHTTPresponsesrecordedduringthetest.ThemostcommonresponsecodeswillbeHTTP200,indicatingsuccess.ForalistofallHTTPresponsecodesandtheirmeaningsseetheattacheddocumentintheglossary.
BacktoContents
RunningVusers
TheRunningVusersgraphdisplaysthenumberofvirtualuserscurrentlyexecutingscriptsinthetestscenario.AVuserperformsasequenceoftasksdefinedinatestscriptuntilthetesthascompleteditsduration,theVuserhasexecutedthescriptforarequirednumberofiterations,orthetestcontrollermanuallystopstheVuser.
Vuserswillusuallybeaddedtothescenarioinwhatiscalleda“gradualramp-up,”unlikethegraphabovewhereallVusersareaddedtothescenariosimultaneously.AfterthenumberofVusersreachesalevelsetinthestresstestplan,thatloadismaintainedforasetdurationbeforeagradualramp-downisinitiated.Theramp-downisusuallyfasterthantheramp-up,becauseremovingusersfromthescenariowillalmostneveradverselyaffectserverperformance.
Wheneveraproblemoccursduringatest,youshouldalwaysbeawareofthenumberofVuserscurrentlyrunninginthescenario.Thepurposeofastresstestistodiagnoseproblemsinanapplicationunderasimulatedload,soinanunstableapplicationitistypicaltoseeadecreaseinsomeperformancemeasurewhenthenumberofVuserspassesacertainlevel.AsuddennegativeshiftinperformanceusuallymeansthenumberofrunningVusershasincreased.
Thegraphabovefollowsatypicaltestschedule,witharamp-upof10usersperminutetoapeakloadof250users,thenthetestwasrunatpeakloadfor10minutes,andaramp-downofapproximately25usersperminute.
BacktoContents
HitsperSecond
TheHitsperSecondgraphshowsthenumberofhitsontheWebserverovertime.Thisgraphcandisplaytheentirescenario,orthelast60,180,600or3600seconds.YoucancomparethisgraphtotheTransactionResponseTimegraphtoseehowthenumberofhitsaffectstransactionperformance.
Anincreaseinhitspersecondindicatesanincreasedloadontheserver,socomparinghitspersecondwithtransactionresponsetimeandthroughputcangiveyouagoodideaofhowtheserverisperformingunderincreasedstress.Whenawebserverisfunctioningcorrectly,hitspersecondwilloftenmirrorthenumberofsuccessfulHTTPresponsecodes.Inthegraphabovethereisaspikeinhitspersecondaround25minutesintothetest.ThiscanbeexplainedbycomparinghitspersecondwiththeresultsofothergraphslikerunningVusers,throughput,andtransactionresponsetime.
BacktoContents
Throughput
TheThroughputgraphshowstheamountofdata,measuredinbytes,transmittedbytheWebserverduringeachsecondofthescenariorun.Throughputismeasuredinbytes/secondandrepresentstheamountofdatatransmittedbythewebserveroverthenetworkatanygiventime.Ideallythroughputshouldincreaseashitspersecondincreases,whichindicatesthatthereissufficientbandwidthtoprocessuserrequests.
YoucangaininformationonserverperformancebycomparingThroughputandHitsperSecond.Iftheserverisperformingasexpectedashitspersecondincrease,thenthroughputshouldalsoincrease.Thisisexpectedbehaviorbecauseanincreaseinhitswillrequiretheservertotransmitmoredatabacktotheusers.However,ifthenumberofhitspersecondincreasesandthroughputstaysconstantordecreases,thissignifiesthattheservercannotprocesstheincreaseinrequests(hitspersecond),resultinginanincreaseintransactionresponsetime.
Thegraphaboveexhibitsageneraldecreaseinthroughputduringtheentiretest,withaslightbutnoticeabledecreasejustbefore25minutes,correspondingtothespikeinhitspersecondinthepreviousgraph.Tocomparethesestatisticsitisbesttooverlaythegraphs.
BacktoContents
HitsperSecond-Throughput
Thisgraphcontainsanoverlayofhitspersecondandthroughput.Wecanseeherethatthespikeinhitspersecondcoincideswithadecreaseinthroughputataround25minutes.Ifaserverisoperatingbelowpeakload,thenhitspersecondusuallymirrorsthroughput.Theresultinthegraphaboveismostlikelyduetoanincreaseinuserload,becausemoreuserswillcauseanincreaseinhitspersecond,andwhentheservercannolongerprocessthenumberofrequestscomingintherewillbeadecreaseinthroughput.
ReferringbacktheRunningVusersgraph,weknowthatthepeakloadof250userswasreachedjustbefore25minutes,resultinginanexpectedincreaseinhitspersecondandadecreaseinthroughput,indicatingthattheloadontheserverwasapproachingitsmaximum.Wecanguessfromthisgraphthattransactionresponsetimeandpossiblyerrorspersecondwillspikearound25minutesaswell.
BacktoContents
AverageTransactionResponseTime
TheAverageTransactionResponsetimegraphshowstheaverageresponsetimeofindividualtransactionsovertime.Responsetimeistheamountoftimebetweenaclientrequestandserverresponse.OftenresponsetimewillcorrespondtothenumberofVuserscurrentlyrunninginthetestscenario.
ThemostcommoncauseforanincreaseinresponsetimeisanincreaseinrunningVusers.ThisoccursbecauseasmoreVusersenterthetestscenariotheamountofworktheservermustperformincreases.Withanincreasedloadittakesmoretimefortheservertorespondtoanindividualrequest,soaverageresponsetimeincreasesaswell.
Inthegraphaboveweshouldexpectanincreaseinresponsetimecorrespondingtothechangeinbehaviorinothergraphsataround25minutes.Lookingattheresponsetimeforthe“Z_”transactionistheeasiestwaytoanalyzegeneralbehavior,andheretotalresponsetimeincreasedtoanaverageofaround110secondsby10minutesintothescenario.At10minutesthenumberofrunningVuserswas120,soitislikelythattheserverwasoperatingathighcapacityfromthistimeonwards.
BacktoContents
TransactionsperSecond(Passed/Failed)
TheTransactionsperSecondgraphshowsthenumberofsuccessful/unsuccessfultransactionsovertime.AtransactionrepresentsasinglesteporgroupofstepsexecutedbyaVuser.Forexample,theprocessofenteringandsubmittinglogininformationonawebpagecouldbegroupedasasingletransaction.
ThenumberoftransactionspersecondshouldcorrespondtothenumberofVusersrunninginthetestscenarioatagiventime,andtypicallyasmoreVusersareaddedtothescenariothenumberoftransactionspersecondwillincrease.IfthenumberoftransactionsperseconddecreaseswithanincreaseinVusers,thenthereisprobablysomepro
- 配套讲稿:
如PPT文件的首页显示word图标,表示该PPT已包含配套word讲稿。双击word图标可打开word文档。
- 特殊限制:
部分文档作品中含有的国旗、国徽等图片,仅作为作品整体效果示例展示,禁止商用。设计者仅对作品中独创性部分享有著作权。
- 关 键 词:
- Loadrunner 结果分析向导1 结果 分析 向导
![提示](https://static.bdocx.com/images/bang_tan.gif)