machine data 101

104
Copyright © 2014 Splunk Inc. Machine Data 101 Gary Burgett Sr. SE 11/1/2016

Upload: splunk

Post on 16-Apr-2017

174 views

Category:

Technology


0 download

TRANSCRIPT

Page 1: Machine Data 101

Copyright©2014SplunkInc.

MachineData101

GaryBurgettSr.SE

11/1/2016

Page 2: Machine Data 101

WhatDoesMachineDataLookLike?Sources

OrderProcessing

Twitter

CareIVR

MiddlewareError

2

Page 3: Machine Data 101

MachineDataContainsCriticalInsightsCustomerID OrderID

Customer’sTweet

TimeWaitingOnHold

TwitterID

ProductID

Company’sTwitterID

CustomerIDOrderID

CustomerID

Sources

OrderProcessing

Twitter

CareIVR

MiddlewareError

3

Page 4: Machine Data 101

MachineDataContainsCriticalInsightsOrderID

Customer’sTweet

TimeWaitingOnHold

ProductID

Company’sTwitterID

OrderID

CustomerID

TwitterID

CustomerID

CustomerID

Sources

OrderProcessing

Twitter

CareIVR

MiddlewareError

4

Page 5: Machine Data 101

StructuredRDBMS

SQL Search

SchemaatWrite SchemaatRead

Traditional Splunk

SplunkApproachtoMachineData

Copyright © 2014 Splunk Inc. 5

ETL UniversalIndexing

Volume Velocity Variety

Unstructured

Page 6: Machine Data 101

Splunk:ThePlatformforMachineData

6

DeveloperPlatform

Reportand

analyze

Customdashboards

Monitorandalert

Adhocsearch

OnlineServices

WebProxy

DataLossPrevention

Storage Desktops

PackagedApplications

CustomApplications

Databases

CallDetailRecords

SmartphonesandDevices

FirewallAuthentication

Fileservers

Endpoint

ThreatIntelligence

Asset&CMDB

Employee/HRInfo

DataStoresApplications

ExternalLookups

Badgingrecords

Emailservers

VPN

Anyamount,anylocation,anysource

Schema-on-the-fly

Universalindexing

Noback-endRDBMS

Noneedtofilterdata

Page 7: Machine Data 101

PlatformforOperationalIntelligence

TheSplunkPortfolio

RichEcosystemofApps&Add-Ons

SplunkPremiumSolutions

MainframeData

RelationalDatabases

MobileForwarders Syslog/TCP IoTDevices

NetworkWireData

Hadoop

Page 8: Machine Data 101

Agenda

§ Non-TraditionalDataSources

§ DataEnrichment

§ LevelUponSearchandReportingCommands

§ DataModelsandPivot

§ AdvancedVisualizationsandtheWebFramework

8

Page 9: Machine Data 101

WorkshopSetup

Page 10: Machine Data 101

Non-TraditionalDataSources

Page 11: Machine Data 101

Non-TraditionalDataSources

§ NetworkInputs

§ HTTPEventCollector

§ LogEventAlertAction

§ SplunkAppforStream

§ ScriptedInputs

§ DatabaseInputs

§ SplunkODBCDriver

§ ModularInputs

§ zLinux Forwarder

§ MINT

§ Non-SplunkDatastores

11

Page 12: Machine Data 101

TraditionalDataSources§ Captureseventsfromlogfilesinrealtime

§ Runsscriptstogathersystemmetrics,connecttoAPIsanddatabases

§ Listenstosyslog andgathersWindowsevents

§ Universallyindexesanydataformatsoitdoesn’tneedadapters

12

Windows• Registry• Eventlogs• Filesystem• sysinternals

Linux/Unix• Configurations• Syslog• Filesystem• Ps,iostat,top

Virtualization• Hypervisor• GuestOS• GuestApps

Applications• Weblogs• Log4J,JMS,JMX• .NETevents• Codeandscripts

Databases• Configurations• Audit/querylogs• Tables• Schemas

Network• Configurations• syslog• SNMP• netflow

Page 13: Machine Data 101

NetworkInputs

§ CollectdataoveranyUDPorTCPport§ Somedevicesonlysenddataoveranetworkport

§ BestPractice:usesyslog-ng orrsyslog§ Offerspersistence§ Categorizesdatabyhost

13

Page 14: Machine Data 101

HTTPEventCollector(HEC)

§ CollectdataoverHTTPorHTTPSdirectlytoSplunk§ ApplicationDeveloperfocus– fewlinesofcodeinapp

tosenddata§ HECFeaturesInclude:

§ Token-based,notcredentialbased§ IndexerAcknowledgements– guaranteesdataindexing§ RawandJSONformattedeventpayloads§ SSL,CORS(CrossOrigion access),andNetworkRestrictions

14

Page 15: Machine Data 101

LogEventAlertAction

§ UseSplunkalertingtoindexacustomlogevent§ Splunksearchableindexofcustomalertevents

§ ConfigurableFeaturesInclude:§ Host§ Source§ Sourcetype§ Index§ Eventtext– constructtheexactsyntaxofthelogevent,

includinganytext,tokens,orotherinformation

15

Page 16: Machine Data 101

TheSplunkAppforStream

WireDataEnhancesthePlatformforOperationalIntelligence

Efficient,Cloud-readyWireDataCollection

SimpleDeploymentSupportsFastTimetoValue

16

Page 17: Machine Data 101

Stream=BetterInsightsfor*

SolutionArea ContextualData WireData Enriched View

ApplicationManagement

applicationlogs,monitoringdata,metrics,events

protocolconversationsondatabaseperformance,DNSlookups,clientdata,businesstransactionpaths…

Measureapplicationresponsetimes,deeperinsightsforroot-causediagnostics,tracetxpaths,establishbaselines…

IT Operations applicationlogs,monitoringdata,metrics,events

payloaddataincludingprocesstimes,errors,transactiontraces,ICAlatency,SQLstatements,DNSrecords…

Analyzetrafficvolume,speedandpacketstoidentifyinfrastructureperformanceissues,capacityconstraints,changes;establishbaselines…

17

Page 18: Machine Data 101

Stream=BetterInsightsfor*SolutionArea ContextualData WireData Enriched View

Security app+infralogs,monitoringdata,events

protocolidentification,protocolheaders,contentandpayloadinformation,flowrecords

Buildanalyticsandcontextforincidentresponse,threatdetection,monitoringandcompliance

DigitalIntelligence

websiteactivity,clickstreamdata,metrics

browser-levelcustomerinteractions

CustomerExperience – analyzewebsiteandapplicationbottleneckstoimprovecustomerexperienceandonlinerevenues

CustomerSupport(online,callcenter)– fasterrootcauseanalysisandresolutionofcustomerissueswithwebsiteorapps

18

Page 19: Machine Data 101

ScriptedInputs

19

§ SenddatatoSplunkviaacustomscript§ Splunkindexesanythingwrittentostdout§ Splunkhandlesscheduling§ Supportsshell,Pythonscripts,WINbatch,PowerShell§ Anyotherutilitythatcanformatandstreamdata

StreamingMode§ Splunkexecutesscriptandindexesstdout

§ Checksforanyrunninginstances

WritetoFileMode§ Splunklaunchesscriptwhichproducesoutputfile,noneedforexternalscheduler

§ Splunkmonitorsoutputfile

Page 20: Machine Data 101

UseCasesforScriptedInputs

20

§ Alternativetofile-baseornetwork-basedinputs§ Streamdatafromcommand-linetools,suchasvmstat andiostat§ Pollawebservice,APIordatabaseandprocesstheresults§ Reformatcomplexorbinarydataforeasierparsingintoeventsandfields§ Maintaindatasourceswithsloworresource-intensivestartup

procedures§ Providespecialorcomplexhandlingfortransientorunstableinputs§ Scriptsthatmanagepasswordsandcredentials§ Wrapperscriptsforcommandlineinputsthatcontainspecialcharacters

Page 21: Machine Data 101

DatabaseInputs

§ Createvaluewithstructureddata§ Enrichsearchresultswithadditionalbusinesscontext

§ Easilyimportdatafordeeperanalysis§ IntegratemultipleDBsconcurrently§ Simpleset-up,non-invasiveandsecure

DBConnectprovidesreliable,scalable,real-timeintegrationbetweenSplunkandtraditionalrelationaldatabases

21

Page 22: Machine Data 101

ConfigureDatabaseInputs

22

§ DBConnectApp§ Real-time,scalableintegrationwithrelationalDBs§ Browseandnavigateschemasandtablesbeforedataimport§ Reliablescheduledimport§ SeamlessinstallationandUIconfiguration§ Supportsconnectionpoolingandcaching

§ “Tail”tablesorimportentiretables§ Detectandimportnew/updatedrowsusingtimestampsoruniqueIDs

§ SupportsmanyRDBMSflavors§ AWSRDSAurora,AWSRedShift,IBMDB2forLinux,Informix,MemSQL,MSSQL,MySQL,

Oracle,PostgreSQL,SAPSQLAnywhere(akaSybaseSA),SybaseASEandIQ,Teradata

Page 23: Machine Data 101

SplunkODBCDriver

23

§ Interactwith,manipulateandvisualizemachinedatainSplunkEnterpriseusingbusinesssoftwaretools

§ LeverageanalyticsfromSplunkalongsideMicrosoftExcel,TableauDesktoporMicrostrategy AnalyticsDesktop

§ Industry-standardconnectivitytoSplunkEnterprise§ Empowersbusinessuserswithdirectandsecureaccesstomachinedata

§ Combinemachinedatawithstructureddataforbetteroperationalcontext

Page 24: Machine Data 101

ODBC:HowitWorks

24

Page 25: Machine Data 101

ModularInputs

25

§ Createyourowncustominputs§ Scriptedinputwithstructureandintelligence§ FirstclasscitizenintheSplunkmanagementinterface§ AppearsunderSettings>DataInputs

§ Benefitsoversimplescriptedinput§ Instancecontrol:launchasingleormultipleinstances§ Inputvalidation§ Supportmultipleplatforms§ StreamdataastextorXML§ SecureaccesstomodinputscriptsviaRESTendpoints

Page 26: Machine Data 101

ExampleModularInputs

26

Twitter§ StreamJSONdatafromaTwittersourcetoSplunkusingTweepy

AmazonS3OnlineStorage§ IndexdatafromtheAmazonS3onlinestoragewebservice

JavaMessagingService(JMS)§ PollmessagequeuesandtopicsthroughJMSMessagingAPI§ Talkstomultipleproviders:MQSeries (Websphere MQ),ActiveMQ,TibcoEMS,HornetQ,RabbitMQ,NativeJMS,WebLogic JMS,SonicMQ

SplunkWindowsInputs§ RetrieveWINeventlogs,registrykeys,perfmon counters

Page 27: Machine Data 101

MoreModularInputs

27

Page 28: Machine Data 101

zLinux Forwarder

28

§ EasilycollectandindexdataonIBMmainframes

§ Collectapplicationandplatformdata

§ DownloadasnewForwarderdistributionfors390xLinux

Page 29: Machine Data 101

ExtendOperationalIntelligencetoMobileApps

29

DeliverBetterPerforming,MoreReliableApps

DeliverReal-TimeOmni-Channel

Analytics

End-to-EndPerformanceandCapacityInsights

Page 30: Machine Data 101

MonitorAppUsageandPerformance

• Improveuserretentionbyquicklyidentifyingcrashesandperformanceissues

• Establishwhetherissuesarecausedbyanapporthenetwork(s)

• Correlateapp,OSanddevicetypetodiagnosecrashandnetworkperformanceissues

30

Page 31: Machine Data 101

IntegratedAnalyticsPlatformforDiverseDataStoresFull-featured,IntegratedProduct

FastInsightsforEveryone

WorkswithWhatYouHaveToday

Explore Visualize Dashboards

ShareAnalyze

HadoopClusters NoSQLandOtherDataStores

Hadoop ClientLibraries StreamingResourceLibraries

Bi-directionalIntegrationwithHadoop

Page 32: Machine Data 101

ConnecttoNoSQLandOtherDataStores

• Buildcustomstreamingresourcelibraries

• SearchandanalyzedatafromotherdatastoresinHunk

• InpartnershipwithleadingNoSQLvendors

• UseinconjunctionwithDBConnectforrelationaldatabaselookups

Page 33: Machine Data 101

VirtualIndexes

§ EnablesseamlessuseofalmosttheentireSplunkstackondata

§ AutomaticallyhandlesMapReduce

§ Technologyispatentpending

Page 34: Machine Data 101

DataEnrichment

Page 35: Machine Data 101

Agenda

§ Tags – categorizeandaddmeaningtodata

§ FieldAliases – simplifysearchandcorrelation

§ CalculatedFields – shortcutcomplex/repetitivecomputations

§ EventTypes – groupcommoneventsandshareknowledge

§ Lookups – augmentdatawithadditionalexternalfields

35

Page 36: Machine Data 101

§ Addsinlinemeaning/context/specificitytorawdata

§ Usedtonormalizemetadataorrawdata

§ Simplifiescorrelationofmultipledatasources

§ CreatedinSplunk

§ Transferredfromexternalsources

WhatisDataEnrichment?

36

Page 37: Machine Data 101

§ Addmeaning/context/specificitytorawdata

§ Labelsdescribingteam,category,platform,geography

§ Appliedtofield-valuecombination

§ Multipletagscanbeappliedforeachfield-value

§ Casesensitive

Tags

37

Page 38: Machine Data 101

CreateTags

38

SHOW

Page 39: Machine Data 101

§ Searcheventswithtaginanyfield

§ Searcheventswithtaginaspecificfield

§ Searcheventswithtagusingwildcards

FindtheWebServersTagsinAction

39

tag=webserver

tag::host=webserver

tag=web*

§ Tagthehostaswebserver

§ Tagthesourcetypeasweb

1

2

3

4

5

SHOW

BacktoSlides

Page 40: Machine Data 101

§ Normalizefieldlabelstosimplifysearchandcorrelation§ Applymultiplealiasestoasinglefield

§ Example:Username|cs_username |Userà user§ Example:c_ip |client|client_ipà clientip

§ Processedafterfieldextractions+beforelookups

§ Canapplytolookups

§ Aliasesappearalongsideoriginalfields

FieldAliases

40

Page 41: Machine Data 101

Re-LabelFieldtoIntuitiveNameCreateFieldAlias

41

1

2

3

SHOW

Page 42: Machine Data 101

§ Createfieldaliasofclientip=customer

§ Searcheventsinlast15minutes,findcustomerfield

§ Fieldalias(customer)andoriginalfield(clientip)arebothdisplayed

SearchusinganIntuitiveFieldNameFieldAliasinAction

42

1

3

2

sourcetype=access_combined

SHOW

Page 43: Machine Data 101

§ Shortcutforperformingrepetitive/long/complextransformationsusingevalcommand

§ Basedonextractedordiscoveredfieldsonly

§ Donotapplytolookuporgeneratedfields

CalculatedFields

43

Page 44: Machine Data 101

ComputeKilobytesfromBytesCreateCalculatedField

44

1

21

2

3

SHOW

Page 45: Machine Data 101

§ Createkilobytes=bytes/1024

§ Searcheventsinlast15minutesforkilobytesandbytes

SearchUsingKilobytesinsteadofBytesCalculatedFieldsinAction

45

1

2

sourcetype=access_combined

SHOW

BacktoSlides

Page 46: Machine Data 101

§ Classifyandgroupcommonevents

§ Captureandshareknowledge

§ Basedonsearch

§ Useincombinationwithfieldsandtagstodefineeventtopography

EventTypes

46

Page 47: Machine Data 101

§ BestPractice:Usepunctfield§ Defaultmetadatafielddescribingeventstructure§ Builtoninterestingcharacters:",;-#$%&+./:=?@\\'|*\n\r\"(){}<>[]^! »§ Canusewildcards

CreateEventTypes

47

event punct

####<Jun3,20145:38:22PMMDT><Notice><WebLogicServer><bea03><asiAdminServer><WrapperStartStopAppMain><>WLSKernel<><><BEA-000360><ServerstartedinRUNNINGmode>

####<_,__::__>_<>_<>_<>_<>_<>_

172.26.34.223- - [01/Jul/2005:12:05:27-0700]"GET/trade/app?action=logoutHTTP/1.1"2002953

..._-_-_[:::_-]_\"_?=_/.\"__

Page 48: Machine Data 101

§ Showpunctforsourcetype=access_combined

§ Pickapunct,thenwildcarditafterthetimestamp

§ AddNOTstatus=200

§ Saveas“bad”eventtype+Color:red+Priority:1(shiftreloadinbrowsertoshowcoloring)

ClassifyEventsasKnownBadCreateEventType

48

eventtype=bad

sourcetype="access_combined" punct="..._-_-_[//_:::]*" NOT status=200

1

2

3

4

SHOW

BacktoSlides

Page 49: Machine Data 101

LookupstoEnrichRawData

LDAPAD

WatchLists

CRM/ERP

CMDB

ExternalDataSources

Insightcomesout

DatagoesinCreateadditionalfieldsfromtherawdatawithalookuptoanexternaldatasource

Page 50: Machine Data 101

§ Augmentraweventswithadditionalfields§ Providecontextorsupportingdetails

§ Translatefieldvaluestomoredescriptivedata§ Example:addtextdescriptionsforerrorcodes,IDs§ Example:addcontactdetailstousernamesorIDs§ Example:adddescriptionstoHTTPstatuscodes

§ File-basedorscriptedlookups

Lookups

50

Page 51: Machine Data 101

51

1.Upload/createtable

2.Assigntabletolookupobject

3.Maplookuptodataset

Convert a Code into a DescriptionConfigure a Static Lookup

SHOW

Page 52: Machine Data 101

§ GetthelookupfromtheSplunkWiki(saveto.csv file)http://wiki.splunk.com/Http_status.csv

§ Lookuptablefiles>Addnew§ Name:http_status.csv (musthave.csv fileextension)§ Upload:<pathto.csv>

§ Verifylookupwascreatedsuccessfully

1.CreateHTTPStatusTable

52

SHOW

| inputlookup http_status.csv

1

2

3

Page 53: Machine Data 101

§ Lookupdefinitions>Addnew§ Name:http_status§ Type:File-based§ Lookupfile:http_status.csv

§ Invokethelookupmanually

2.AddLookupDefinition

53

SHOW

1

2

sourcetype=access_combined | lookup http_status status OUTPUT status_description

Page 54: Machine Data 101

§ Automaticlookups>Addnew§ Name:http_status (cannothavespaces)§ Lookuptable:http_status§ Applyto:sourcetype=access_combined§ Lookupinputfield:status§ Lookupoutputfield:status_description

§ Verifylookupisinvokedautomatically

3.ConfigureAutomaticLookup

54

SHOW

1

2

sourcetype=access_combinedBacktoSlides

Page 55: Machine Data 101

§ Temporallookupsfortime-basedlookups§ Example:IdentifyusersonyournetworkbasedontheirIPaddress

andthetimestampinDHCPlogs

§ Usesearchresultstopopulatealookuptable§ … | outputlookup <tablename|filename>

§ Callanexternalcommandorscript§ Pythonscriptsonly§ Example:DNSlookupforIPßà Host

§ Createalookuptableusingarelationaldatabase§ ReviewmatchesagainstadatabasecolumnorSQLquery

FancyLookups

55

Page 56: Machine Data 101

§ CreatingandManagingAlerts(JobInspector)

§ Macros

§ WorkflowActions

MoreDataEnrichment

56

Page 57: Machine Data 101

LevelUponSearch&ReportingCommands

Page 58: Machine Data 101

Agenda

§ Doingmorewithbasicsearchcommands

§ Advancedsearchcommands

§ Doingmorewithbasicreportingcommands

58

Page 59: Machine Data 101

SearchSyntaxComponents

59

Page 60: Machine Data 101

AnatomyofaSearch

60

Disk

Page 61: Machine Data 101

§ top– limit§ rare– sameoptionsastop§ timechart– parameters§ stats– functions(sum,avg,list,values,sparkline)§ sort– inlineascendingordescending§ addcoltotals§ addtotals

DoingMorewithBasicSearchCommands

61

Page 62: Machine Data 101

WorkshopNotesforPresenter

Tip#5:Inthenextsection,aftereachsearch,havetheparticipantssavethesearchasadashboardpanel.Attheend

oftheworkshop,theywillhavealivingdocumentoftheworkshopexercisestoreferencelater.

Acompleteversionofthisdashboardispackagedasanapp.ItisuploadedtotheBoxfolderasaleavebehind.

62

Page 63: Machine Data 101

§ Commandshaveparametersorqualifiers

§ topandrarehavesimilarsyntax

§ Eachsearchcommandhasitsownsyntax– showinlinehelp

FindMostandLeastActiveCustomersUsingthetop+rareCommands

... | top limit=20 clientip

... | rare limit=20 clientip

IPswiththemostvisits

IPswiththeleastvisits

SHOW

Page 64: Machine Data 101

§ Sortinlinedescendingorascending

64

... | stats count by clientip | sort - count

... | stats count by clientip | sort + count

Numberofrequestsbycustomer- descending

Numberofrequestsbycustomer- ascending

SorttheNumberofCustomerRequestsUsingthesortCommand

SHOW

Page 65: Machine Data 101

§ ShowSearchCommandReferenceDocs§ Functionsforeval+where§ Functionsforstats+chartandtimechart

§ Invokeafunction

§ Renameinline

65

... | stats sum(bytes) by clientip | sort - sum(bytes)

... | stats sum(bytes) as totalbytes by clientip | sort - totalbytes

Totalpayloadbycustomer- descending

Totalpayloadbycustomer- ascending

DetermineTotalCustomerPayloadUsingfunctions+renamecommand

SHOW

Page 66: Machine Data 101

§ Listallvaluesofafield

§ Listonlydistinctvaluesofafield

66

... | stats values(action) by clientip

... | stats list(action) by clientip

Activitybycustomer

Distinctactionsbycustomer

ObserveCustomerActivityUsingthelist+valuesFunctions

SHOW

Page 67: Machine Data 101

§ Showdistinctactionsandcardinalityofeachaction

67

sourcetype=access_combined| stats count(action) as value by clientip, action| eval pair=action + " (" + value + ")"| stats list(pair) as values by clientip

AnalyzeCustomerActivityCombinelist+valuesFunctions

SHOW

Page 68: Machine Data 101

§ Addcolumns

§ Sumspecificcolumns

68

... | stats count by clientip, action

2cols:clientip +action

... | stats sum(bytes) as totalbytes, avg(bytes) as avgbytes, count as totalevents by clientip | addcoltotals totalbytes, totalevents

Sumtotalbytesandtotaleventscolums

BuildingaTableofCustomerActivityAddColumnsandSumColumns

SHOW

Page 69: Machine Data 101

69

... | stats sum(bytes) as totalbytes, sum(other) as totalother by clientip | addtotals fieldname=totalstuff

Foreachrow,addtotalbytes+totalother

Abetterexample:physicalmemory+virtualmemory=

totalmemory

BuildingaTableofCustomerActivitySumAcrossRows

SHOW

Page 70: Machine Data 101

70

... | stats sparkline(count) as trendline by clientip

Incontextoflargereventset

... | stats sparkline(count) as trendline sum(bytes) by clientip

Inlineintables

TrendIndividualCustomerActivitySparklinesinAction

SHOW

BacktoSlides

Page 71: Machine Data 101

AdvancedSearchCommandsCommand ShortDescription Hints

transaction Groupeventsbyacommonfieldvalue. Convenient,but resourceintensive.cluster Clustersimilareventstogether. Canbeusedon_raworfield.associate Identifiescorrelationsbetweenfields. Calculatesentropybtn fieldvalues.correlate Calculatesthecorrelationbetween

differentfields.Evaluatesrelationshipof allfieldsinaresultset.

contingency Buildsacontingencytablefortwofields. Computesco-occurrence,or%twofieldsexistinsameevents.

anomalies Computesanunexpectednessscoreforanevent.

Computessimilarityofevent(X)toasetofpreviousevents(P).

anomalousvalue Findsandsummarizesirregular,oruncommon,searchresults.

Considers frequencyofoccurrenceornumberofstdev fromthemean

Page 72: Machine Data 101

§ Seweventstogether+createsduration+eventcount

§ Sparklinesinlineintables

72

... | transaction JSESSIONID | table JSESSIONID, action, product_id

GroupbyJSESSIONID

ViewCustomerActivitybySessionUsingthetransactionCommand

SHOW

Page 73: Machine Data 101

§ Intelligentgroup(createscluster_countandcluster_label)

§ Sparklinesinlineintables

Cluster

73

SHOW

... | cluster showcount=1 | table _raw, cluster_count, cluster_label

BacktoSlides

Page 74: Machine Data 101

§ Predictovertime

§ ChartOverlaywithandwithoutstreamstats

§ Mapswithiplocation+geostats

§ Singlevalue

§ Meteredvisualswithgauge

DoingMorewithBasicReportingCommands

74

Page 75: Machine Data 101

§ Predictfuturevaluesusinglower/upperbounds– singleandmultipleseries

75

... | timechart count as traffic | predict traffic

PredictWebsiteTrafficUsingthepredictCommand

SHOW

Page 76: Machine Data 101

76

sourcetype=access_combined (action=view OR action=purchase)| timechart span=10m count(eval(action="view")) as Viewed,

count(eval(action="purchase")) as Purchased

CompareBrowsingvs.BuyingActivitySimpleChartOverlay

SHOW

Page 77: Machine Data 101

77

... | iplocation clientip | geostats count by clientip

CombineIPlookupwithgeomapping

MapCustomerActivity GeographicallyGeolocation inAction

SHOW

Page 78: Machine Data 101

78

... | stats count

DisplayaSimpleCountofEventsSingleValueinAction

SHOW

Page 79: Machine Data 101

DisplayCountsUsingGaugesSingleValue,RadialandFillerGaugesinAction

79

... | stats count | gauge count 10000 20000 30000 40000 50000

SHOW

BacktoSlides

Page 80: Machine Data 101

DataModelandPivot

Page 81: Machine Data 101

Agenda

§ Whatisadatamodel?

§ Buildadatamodel

§ PivotInterface

§ Accelerateadatamodel

81

Page 82: Machine Data 101

PowerfulAnalyticsAnyoneCanUse

Enablesnon-technicaluserstobuildcomplexreportswithoutthesearchlanguage

Providesmoremeaningfulrepresentationofunderlyingrawmachinedata

Accelerationtechnologydeliversupto1000xfasteranalyticsoverSplunk5

82

Pivot

DataModel

AnalyticsStore

Page 83: Machine Data 101

DefineRelationshipsinMachineDataDataModel• Describeshowunderlyingmachinedataisrepresentedandaccessed

• Definesmeaningfulrelationshipsinthedata

• Enablessingleauthoritativeviewofunderlyingrawdata

Hierarchicalobjectviewofunderlyingdata

Addconstraintstofilteroutevents

Page 84: Machine Data 101

TransparentAcceleration

• Automaticallycollected– Handlestimingissues,

backfill…• Automaticallymaintained– Usesaccelerationwindow

• Storedontheindexers– Peertothebuckets

• Faulttolerantcollection

Timewindowofdatathatisaccelerated

Checktoenableaccelerationofdatamodel

HighPerformanceAnalyticsStore

Page 85: Machine Data 101

Easy-to-UseAnalytics

• Drag-and-dropinterfaceenablesanyusertoanalyzedata

• Createcomplexqueriesandreportswithoutlearningsearchlanguage

• Clicktovisualizeanycharttype;reportsdynamicallyupdatewhenfieldschange

Selectfieldsfromdatamodel

Timewindow

Allcharttypesavailableinthecharttoolbox

Savereporttoshare

Pivot

Page 86: Machine Data 101

§ Definesleastcommondenominatorforadatadomain

§ Standardmethodtoparse,categorize,normalizedata

§ Setoffieldnamesandtagsbydomain§ PackagedasaDataModelsinaSplunkApp

§ Domains:security,web,inventory,JVM,performance,networksessions,andmore

§ MinimalsetuptousePivotinterface

CommonInformationModel(CIM)App

86

Page 87: Machine Data 101

§ Apps>FindMoreApps>

§ Search:“CommonInformationModel”

§ Installfree

§ Showfieldsforweb+WebDataModel

DownloadCIMApp

87

SHOW

1

2

3

4

BacktoSlides

Page 88: Machine Data 101

DataModel&PivotTutorial

http://docs.splunk.com/Documentation/Splunk/latest/PivotTuto

rial/WelcometothePivotTutorial

88

Page 89: Machine Data 101

CustomVisualizationsandtheWebFrameworkToolkit

Page 90: Machine Data 101

Agenda

§ DeveloperPlatform

§ WebFrameworkToolkit(WFT)

§ RESTAPIandSDKs

§ GetaFlyingStart

90

Page 91: Machine Data 101

OptimizingtheAnalyticsProcess

91

Focusonthedata– intuitivetoolstoenabletheanalyst

Nosinglevisualizationexiststohandlealldatasets.

Neverlosesightoftherawdata

SplunkAnalytics

Explore

Context

Visualize

Algorithms

Page 92: Machine Data 101

6.0+6.1:Simple,Interactive,andExtensible

92

VISUALIZATIONEXPLORATION

CUSTOMIZABLEFRAMEWORK

POWERFULANALYTICS

PivotDataModels

InteractiveFormsContextualDrilldown

DashboardEditorWebFramework

Page 93: Machine Data 101

TheSplunkEnterprisePlatform

Collection

Indexing

SearchProcessingLanguage

CoreFunctions

Inputs,Apps,OtherContent

SDKContent

CoreEngine

UserandDeveloperInterfaces

WebFramework

RESTAPI

Page 94: Machine Data 101

What’sPossiblewiththeSplunkEnterprisePlatform?

PowerMobileApps

LogDirectly

ExtractData

CustomerDashboards

IntegrateBITools

IntegratePlatformServices

Developer Platform

Page 95: Machine Data 101

PowerfulPlatformforEnterpriseDevelopersDevelopers Can Customize and Extend

RESTAPI

BuildSplunkApps ExtendandIntegrateSplunk

SimpleXML

JavaScript

HTML5

WebFramework

JavaJavaScriptPython

RubyC#PHP

DataModels

SearchExtensibility

ModularInputs

SDKs

Page 96: Machine Data 101

SplunkSoftwareforDevelopers

GainApplicationIntelligence

BuildSplunkApps

IntegrateandExtendSplunk

Page 97: Machine Data 101

AWealthofSplunk AppsOver1,100appsavailableontheSplunkappssite

APISDKs UI

Server, Storage, Network

Server Virtualization

Operating Systems

Custom Applications

Business Applications Cloud Services

App Performance MonitoringTicketing/ and Other

WebIntelligence

Mobile Applications

Stream

Page 98: Machine Data 101

§ Interactive,cut/pasteexamplesfrompopularsourcerepositories:D3,GitHub,jQuery

§ Splunk6.xDashboardExamplesApphttps://apps.splunk.com/app/1603

§ CustomSimpleXML ExtensionsApphttps://apps.splunk.com/app/1772

§ SplunkWebFrameworkToolkitApphttps://apps.splunk.com/app/1613

ExampleAdvancedVisualizations

98

Page 99: Machine Data 101

99

http://www.d3js.org

Page 100: Machine Data 101

AddaD3BubbleChart

100

1. GotoFindMoreAppsandInstalltheSplunk6.xDashboardExamplesApp

2. EntertheApp3. GotoExamples>CustomVisualizations>

D3BubbleChart4. Copyautodiscover.js (file)+components/bubblechart (dir)

from:$SH/etc/apps/simple_xml_examples/appserver/staticto:$SH/apps/search/appserver/static

5. CopyandpastesimpleXMLtonewdashboard

SHOW

BacktoSlides

Page 101: Machine Data 101

Resources

Page 102: Machine Data 101

SplunkDocumentation

102

• http://docs.splunk.com• OfficialProductDocs• Wikiandcommunitytopics• Updateddaily• Canbeprintedto.PDF

Page 103: Machine Data 101

SplunkAnswers

103

• http://answers.splunk.com• Communitydriven• Splunksupported• Knowledgeexchange• Q&A

Page 104: Machine Data 101

SplunkEducation

104

• RecommendedforUsers– UsingSplunk– Searching&Reporting

• RecommendedforUI/DashboardDevelopers– DevelopingApps

• Instructor-LedCourses– Web– Onsite