UDavid O'Brien kutshanje uphehlelele eyakhe inkampani, iXirus (https://xirus.com.au), egxile kwiimveliso zelifu likaMicrosoft Azure. Ziyilelwe ukuba zihlale zakha kwaye ziqhube izicelo ezixubileyo kumaziko edatha, iindawo ezisecaleni, iiofisi ezikude, kunye nelifu.
UDavid uqeqesha abantu kunye neenkampani kuzo zonke izinto zeMicrosoft Azure kunye ne-Azure DevOps (eyayifudula i-VSTS) kwaye isasebenza ngokubonisana kunye ne-infracoding. Uye waba yiMicrosoft MVP (Microsoft Most Valuable Professional) Imbasa iminyaka emi-5 kwaye kutshanje ufumene imbasa ye-Azure MVP. Njengomququzeleli we-Melbourne Microsoft Cloud kunye ne-Datacentre Meetup, u-O'Brien uthetha rhoqo kwiinkomfa zamazwe ngamazwe, edibanisa umdla wakhe wokuhamba ngehlabathi kunye nomdla wokwabelana ngamabali e-IT noluntu. Ibhlog kaDavid ifumaneka , ukwapapasha uqeqesho lwakhe lwe-intanethi kwi-Pluralsight.
Intetho ithetha ngokubaluleka kweemetrics ekuqondeni okwenzekayo kwindawo okuyo kunye nendlela isicelo sakho esisebenza ngayo. IMicrosoft Azure inendlela enamandla nelula yokubonisa iimetrics kuzo zonke iintlobo zomthwalo onzima, kwaye intetho icacisa ukuba ungazisebenzisa njani zonke.
Ngentsimbi yesi-3 kusasa ngeCawe, ngelixa uleleyo, uvuswe ngumyalezo wombhalo othi: βUsetyenziso oluphezulu aluphenduli kwakhona.β Kuqhubekani? Uphi kwaye yintoni isizathu "seziqhoboshi"? Kule ntetho, uya kufunda malunga neenkonzo iMicrosoft Azure ezibonelela ngayo abathengi ukuba baqokelele iilog kwaye, ngakumbi, iimetrics ukusuka kumthwalo wakho wamafu. UDavid uya kukuxelela ukuba zeziphi iimetriki onokuthi ube nomdla kuzo xa usebenza kwiqonga lelifu kunye nendlela yokufika kuzo. Uya kufunda malunga nezixhobo zomthombo ovulekileyo kunye nokwakhiwa kwedeshibhodi, kwaye uphele unolwazi olwaneleyo ukwenza iideshibhodi zakho.
Kwaye ukuba uvuswa kwakhona ngo-3 ekuseni ngumyalezo wokuba isicelo esibalulekileyo siphazamisekile, unokukhawuleza ufumanise isizathu saso.
Molo, namhlanje siza kuthetha ngeemetrics. Igama lam ndinguDavid O'Brien, ndingumseki kunye nomnini wenkampani encinci yokucebisa yase-Australia, iXirus. Ndiyabulela kwakhona ngokuza apha uzochitha ixesha lakho kunye nam. Ngoko kutheni silapha? Ukuthetha ngeemetrics, okanye kunoko, ndiza kukuxelela ngazo, kwaye ngaphambi kokuba senze naziphi na izinto, masiqale ngethiyori.
Ndiza kukuxelela ukuba zeziphi iimetrics, onokuthi wenze ntoni ngazo, into ekufuneka uyithathele ingqalelo, ukuqokelela njani kwaye wenze ingqokelela yeemetrics eAzure, kwaye yintoni ukubonwa kweemetrics. Ndiza kukubonisa ukuba ezi zinto zikhangeleka njani kwilifu likaMicrosoft kunye nendlela yokusebenza nelilifu.
Ngaphambi kokuba siqale, ndiza kubuza umboniso wezandla kwabo basebenzisa iMicrosoft Azure. Ngubani osebenza nge-AWS? Ndibona abambalwa. Kuthekani ngoGoogle? ALI Cloud? Indoda enye! Kakhulu. Ke zithini iimetrikhi? Inkcazo esemthethweni yeZiko leSizwe leMigangatho kunye neTekhnoloji yase-US ithi: "I-metric ngumgangatho womlinganiselo ochaza iimeko kunye nemigaqo yokulinganisa ipropati kwaye isebenze ukuqonda iziphumo zokulinganisa." Ingaba ithetha ntoni?
Makhe sithathe umzekelo we-metric yokutshintsha indawo yedisk yasimahla yomatshini wenyani. Ngokomzekelo, sinikwa inombolo ye-90, kwaye eli nani lithetha ipesenti, oko kukuthi, inani lendawo yediski yamahhala yi-90%. Ndiyaqaphela ukuba akunomdla kakhulu ukufunda inkcazo yenkcazo yee-metrics, ethatha amaphepha angama-40 kwifomathi ye-pdf.
Nangona kunjalo, i-metric ayithethi ukuba isiphumo somlinganiselo sifunyenwe njani, ibonisa kuphela esi siphumo. Senza ntoni ngeemetrics?
Okokuqala, silinganisa ixabiso lento ukuze sisebenzise isiphumo somlinganiselo.
Ngokomzekelo, sifumene isixa sediski samahhala kwaye ngoku sinokuyisebenzisa, sebenzisa le nkumbulo, njl. Sakuba sifumene iziphumo zemetric, kufuneka sizitolike. Ngokomzekelo, i-metric ibuyisele isiphumo se-90. Sidinga ukwazi ukuba le nombolo ithetha ukuthini: ubungakanani bendawo yamahhala okanye inani lendawo yediski esetyenzisiweyo kwipesenti okanye i-gigabytes, i-network latency ilingana ne-90 ms, njalo njalo, oko kukuthi. , kufuneka sitolike intsingiselo yexabiso lemetriki. Ukuze iimetriki zibe nentsingiselo konke konke, emva kokutolika ixabiso lemetriki enye, kufuneka siqinisekise ukuba amaxabiso amaninzi ayaqokelelwa. Oku kubaluleke kakhulu kuba abantu abaninzi abayazi imfuneko yokuqokelela iimetrics. UMicrosoft wenze kwalula kakhulu ukuqokelela iimetrics, kodwa kukuwe ukuba uqinisekise ukuba ziqokelelwe. Ezi metrics zigcinwa kuphela iintsuku ezingama-41 kwaye zinyamalale ngomhla wama-42. Ngoko ke, kuxhomekeke kwiipropati zezixhobo zakho zangaphandle okanye zangaphakathi, kufuneka unakekele indlela yokugcina i-metrics ngaphezu kweentsuku ze-41 - ngendlela yezigodo, iilogi, njl. Ke, emva kokuqokelela, kuya kufuneka uzibeke kwindawo ethile ekuvumela ukuba ukhuphe zonke izibalo zotshintsho kwiziphumo zemetriki ukuba kukho imfuneko. Nje ukuba uzibeke apho, unokuqalisa ukusebenza nazo ngokufanelekileyo.
Kuphela emva kokuba ufumene iimethrikhi, uzitolike kwaye uziqokelele, unako ukwenza isivumelwano somgangatho wenkonzo ye-SLA. Le SLA isenokungabalulekanga kangako kubathengi bakho; kubaluleke kakhulu koogxa bakho, abaphathi, abo bagcina inkqubo kwaye banenkxalabo malunga nokusebenza kwayo. I-metric inokulinganisa inani lamatikiti - umzekelo, ufumana amathikithi ama-5 ngosuku, kwaye kule meko ibonisa isantya sokuphendula kwizicelo zabasebenzisi kunye nesantya sokusombulula ingxaki. I-metric akufuneki ithi indawo yakho ilayisha nge-20ms okanye isantya sempendulo yakho yi-20ms, i-metric ingaphezulu kwesalathisi sobugcisa esinye.
Ke ngoko, umsebenzi wencoko yethu kukukubonisa ngomfanekiso ocacileyo wondoqo weemethrikhi. I-metric isebenza ukuze ngokujonga kuyo ufumane umfanekiso opheleleyo wenkqubo.
Nje ukuba sibe ne-metric, sinokuqinisekisa nge-99% ukuba inkqubo iyasebenza, kuba ayijongi nje kwifayile yelog ethi inkqubo iyasebenza. Isiqinisekiso se-99% sokugqibela sithetha ukuba, umzekelo, i-99% yexesha i-API iphendula ngesantya esiqhelekileyo se-30 ms. Yile nto kanye enomdla kubasebenzisi bakho, oogxa bakho kunye nabaphathi. Uninzi lwabathengi bethu bajonga iilog zeseva yewebhu, kodwa abaqapheli naziphi na iimpazamo kubo kwaye bacinga ukuba yonke into ilungile. Ngokomzekelo, babona isantya senethiwekhi ye-200 Mb / s kwaye bacinga: "ok, yonke into ilungile!" Kodwa ukufezekisa ezi 200, abasebenzisi bafuna isantya sokuphendula se-30 milliseconds, kwaye oku ngokuchanekileyo isalathisi esingalinganiswanga kwaye singaqokelelwa kwiifayile zelogi. Kwangaxeshanye, abasebenzisi bayamangaliswa kukuba isiza silayisha kancinci, kuba, singenazo iimethrikhi eziyimfuneko, abazazi izizathu zokuziphatha.
Kodwa ekubeni sine-100% uptime SLA, abathengi baqala ukukhalaza kuba indawo eneneni inzima kakhulu ukuyisebenzisa. Ngoko ke, ukudala i-SLA enenjongo, kuyimfuneko ukubona umfanekiso opheleleyo wenkqubo eyenziwe ngamatriki aqokelelweyo. Lo ngumba oqhubekayo endinawo nabanye ababoneleli abathi, xa bedala ii-SLA, bangazi ukuba igama elithi "uptime" lithetha ntoni kwaye kwiimeko ezininzi abachazi kubathengi babo ukuba i-API yabo isebenza njani.
Ukuba udale inkonzo, umzekelo, i-API yomntu wesithathu, kufuneka uqonde ukuba yintoni i-metric yesiphumo se-39,5 ithetha - impendulo, impendulo eyimpumelelo, impendulo kwi-speed ye-20 okanye kwi-5 ms isivinini. Kukuwe ukuba ulungelelanise i-SLA yabo kwi-SLA yakho, kwiimethrikhi zakho.
Nje ukuba uyiqiqe yonke le nto, ungaqala ukwenza ideshibhodi emangalisayo. Ndixelele, ngaba ukhona umntu osele esebenzise isicelo sokubonwa kweGrafana? Kakhulu! Ndingumlandeli omkhulu walo mthombo uvulekileyo kuba le nto isimahla kwaye kulula ukuyisebenzisa.
Ukuba awukayisebenzisi iGrafana, ndiza kukuxelela indlela yokusebenzisana nayo. Nabani na ozelwe kwi-80s kunye ne-90s mhlawumbi ukhumbula i-CareBears? Andazi ukuba zazidume kangakanani ezi bhere eRashiya, kodwa xa kufikwa kwiimethrikhi, kufuneka sibe βziibhereβ ezifanayo. Njengoko benditshilo, udinga umfanekiso omkhulu wendlela yonke inkqubo esebenza ngayo, kwaye ayifanelanga ukuba ibe malunga ne-API yakho, iwebhusayithi yakho, okanye inkonzo esebenza kumatshini obonakalayo.
Kuya kufuneka uququzelele ukuqokelelwa kwezo metrics ezibonisa ngokupheleleyo ukusebenza kwenkqubo yonke. Uninzi lwenu ngabaphuhlisi besoftware, ngoko ke ubomi benu buhlala buguquka, buziqhelanisa neemfuneko zemveliso entsha, kwaye kanye njengoko uxhalabile ngeenkqubo zokubhala iikhowudi, kufuneka uzixhalabise ngeemethrikhi. Kufuneka wazi ukuba i-metric inxulumana njani nomgca ngamnye wekhowudi oyibhalayo. Ngokomzekelo, ngeveki ezayo uqala umkhankaso omtsha wokuthengisa kwaye ulindele inani elikhulu labasebenzisi ukuba batyelele indawo yakho. Ukuhlalutya esi siganeko, uya kufuna i-metrics, kwaye unokufuna ideshibhodi epheleleyo ukulandelela umsebenzi waba bantu. Uya kufuna iimetrics ukuze uqonde ukuba liphumelele kangakanani iphulo lakho lokuthengisa kunye nendlela eliqhuba ngayo. Baza kukunceda, umzekelo, ukuphuhlisa iCRM esebenzayo - inkqubo yokulawula ubudlelwane bomthengi.
Ke masiqalise ngenkonzo yethu yelifu yeAzure. Kulula kakhulu ukufumana kwaye ulungelelanise ingqokelela yeemetrics kuba ineAzure Monitor. Oku kubeka iliso kwindawo eyodwa yolawulo loqwalaselo lwakho. Into nganye ye-Azure ofuna ukuyifaka kwinkqubo yakho ineemethrikhi ezininzi ezenziwe ngokuzenzekelayo. Esi sisicelo sasimahla esisebenza kanye ngaphandle kwebhokisi kwaye asifuni naziphi na iisetingi zangaphambili; awudingi ukubhala okanye "ukukrola" nantoni na kwinkqubo yakho. Siza kukuqinisekisa oku ngokujonga le demo ilandelayo.
Ukongezelela, kunokwenzeka ukuthumela ezi metrics kwizicelo zeqela lesithathu, ezifana ne-Splunk log storage kunye nenkqubo yokuhlalutya, isicelo solawulo lwelog esekelwe kwifu i-SumoLogic, i-ELK log processing tool, kunye ne-IBM Radar. Enyanisweni, kukho ukungafani okuncinci okuxhomekeke kwizixhobo ozisebenzisayo - umatshini obonakalayo, iinkonzo zenethiwekhi, i-database ye-Azure SQL, oko kukuthi, ukusetyenziswa kweemetriki kuyahluka ngokuxhomekeke kwimisebenzi yendawo yakho yokusebenza. Andiyi kuthetha ukuba ezi zintlukwano zinzulu, kodwa, ngelishwa, zisekho, kwaye oku kufuneka kuthathelwe ingqalelo. Ukuvumela kunye nokuthumela i-metrics kunokwenzeka ngeendlela ezininzi: ngePortal, i-CLI/iShell yaMandla, okanye ukusebenzisa iitemplates ze-ARM.
Phambi kokuba ndiqale idemo yam yokuqala, ndiza kuphendula nayiphi na imibuzo onokuba unayo. Ukuba akukho mibuzo, masiqalise. Isikrini sibonisa ukuba libukeka njani iphepha le-Azure Monitor. Ngaba nabani na kuni onokuthi eli sweni alisebenzi?
Ke ngoku yonke into ilungile, unokubona ukuba iinkonzo zemonitha zijongeka njani. Ndingatsho ukuba esi sisixhobo esihle kakhulu kwaye esilula kakhulu somsebenzi wemihla ngemihla. Ingasetyenziselwa ukubeka esweni izicelo, uthungelwano kunye neziseko. Kungekudala, i-interface yokubeka iliso iphuculwe, kwaye ukuba iinkonzo zangaphambili zibekwe kwiindawo ezahlukeneyo, ngoku zonke iinkcukacha kwiinkonzo zihlanganiswe kwiphepha lasekhaya le-monitor.
Itheyibhile yeemetrics yithebhu ecaleni kwendlela yeHomeMonitorMetrics, onokuya kuyo ukuze ubone zonke iimetriki ezikhoyo kwaye ukhethe ezo uzifunayo. Kodwa ukuba ufuna ukwenza ingqokelela yeemetrics, kufuneka usebenzise i-HomeMonitorDiagnostic useto lwendlela yolawulo kwaye ujonge iibhokisi zokukhangela Iimetrics ezinikwe amandla/akhubazekileyo. Ngokungagqibekanga, phantse zonke iimethrikhi zenziwe ukuba zisebenze, kodwa ukuba ufuna ukwenza into eyongezelelweyo, kuya kufuneka utshintshe imeko yokuxilonga ukusuka kuKhubazekile ukuya kuVunyiwe.
Ukwenza oku, nqakraza kumgca wemetric ekhethiweyo kwaye kwithebhu evulayo, vula indlela yoxilongo. Ukuba uya kuhlalutya i-metric ekhethiweyo, ngoko emva kokuchofoza Vula ikhonkco lokuxilonga, kufuneka ukhangele i-Thumela kwi-Log Analytics checkbox kwifestile evelayo.
I-Log Analytics iyafana neSplunk, kodwa ixabisa kancinci. Le nkonzo ikuvumela ukuba uqokelele zonke iimethrikhi zakho, iilogi kunye nayo yonke enye into oyifunayo kwaye uyibeke kwindawo yokusebenza ye-Log Analytics. Le nkonzo isebenzisa ulwimi olukhethekileyo lwe-KQL lokuphendula imibuzo-Kusto Quarry Language, siya kujonga umsebenzi wayo kwidemo elandelayo. Okwangoku, ndiza kuqaphela ukuba ngoncedo lwayo ungadala imibuzo malunga neemetrics, logs, imiqathango, iintsingiselo, iipateni, njl. kwaye wenze iideshibhodi.
Ke, sijonga i-Thumela kwi-Log Analytics checkbox kunye ne-LOG iibhokisi zokukhangela zepaneli: I-DataPlaneRequests, i-MongoRequests kunye ne-QueryRuntimeStatistics, kwaye ngezantsi kwiphaneli ye-METRIC - i-Izicelo zebhokisi yokukhangela. Emva koko sinika igama kwaye sigcine useto. Kumgca womyalelo, oku kubonisa imigca emibini yekhowudi. Ngendlela, iqokobhe le-Azure Cloud ngale ndlela ifana neGoogle, ekuvumela ukuba usebenzise umgca womyalelo kwisiphequluli sakho sewebhu. I-AWS ayinayo into enjalo, ke iAzure ilunge ngakumbi ngale ndlela.
Umzekelo, ndingaqhuba idemo ngokusebenzisa ujongano lwewebhu ngaphandle kokusebenzisa nayiphi na ikhowudi kwilaptop yam. Ukwenza oku, kufuneka ndiqinisekise ngeakhawunti yam yeAzure. Emva koko ungasebenzisa, umzekelo, i-terrafone, ukuba sele uyisebenzisa, linda uqhagamshelo kwinkonzo kwaye ufumane indawo yokusebenza yeLinux esetyenziswa nguMicrosoft ngokungagqibekanga.
Emva koko, ndisebenzisa iBash, eyakhelwe kwi-Azure Cloud Shell. Into eluncedo kakhulu yi-IDE eyakhelwe kwisikhangeli, uguqulelo olulula lweKhowudi yeVS. Okulandelayo, ndingangena kwitemplate yam yeemetrics zempazamo, ndiyihlele, kwaye ndiyenze ngokwezifiso ukuze ihambelane neemfuno zam.
Nje ukuba usete ingqokelela yeemetrics kule template, ungayisebenzisa ukwenza iimetrics zesiseko sakho sonke. Nje ukuba sisebenzise iimetriki, siziqokelele, kwaye sizigcine, kuya kufuneka sizijonge.
I-Azure Monitor isebenza kuphela ngeemetrics kwaye ayiboneleli ngomfanekiso opheleleyo wempilo yenkqubo yakho. Unokuba nenani lezinye izicelo ezisebenza ngaphandle kwendawo yeAzure. Ke ukuba ufuna ukubeka esweni zonke iinkqubo, ukujonga zonke iimetriki eziqokelelweyo kwindawo enye, ke i-Azure Monitor ayifanelekanga oku.
Ukusombulula le ngxaki, iMicrosoft inikezela ngesixhobo se-Power BI, isoftware ebanzi yohlalutyo lweshishini olubandakanya ukubonwa kwedatha eyahlukeneyo. Le yimveliso ebiza kakhulu, ixabiso elixhomekeke kwiseti yemisebenzi oyifunayo. Ngokungagqibekanga, inikezela ngeentlobo ezingama-48 zedatha ekufuneka uyiqhube kwaye inxulunyaniswe neAzure SQL Data Warehouses, Azure Data Lake Storage, Azure Machine Learning Services, kunye neAzure Databricks. Ukusebenzisa i-scalability, unokufumana idatha entsha rhoqo ngemizuzu engama-30. Oku kunokwanela okanye kunganeli kwiimfuno zakho ukuba ufuna ukubonwa kwexesha lokwenyani. Kule meko, kuyacetyiswa ukuba kusetyenziswe izicelo ezifana neGrafana endiyikhankanyileyo. Ukongezelela, amaxwebhu eMicrosoft achaza amandla okuthumela iimethrikhi, iilogi kunye neetafile zesiganeko usebenzisa izixhobo ze-SIEM kwiinkqubo zokubonisa i-Splunk, i-SumoLogic, i-ELK kunye ne-IBM radar.
23:40 imiz
Iza kuqhubeka kungekudala...
Ezinye iintengiso π
Enkosi ngokuhlala nathi. Ngaba uyawathanda amanqaku ethu? Ngaba ufuna ukubona umxholo onomdla ngakumbi? Sixhase ngokufaka iodolo okanye ngokucebisa abahlobo, , i-analogue eyodwa yeeseva zomgangatho wokungena, eyenzelwe wena: (ifumaneka nge-RAID1 kunye ne-RAID10, ukuya kuthi ga kwi-24 cores kunye ne-40GB DDR4).
Dell R730xd 2x ngexabiso eliphantsi kwiziko ledatha le-Equinix Tier IV eAmsterdam? Kuphela apha eNetherlands! Dell R420 - 2x E5-2430 2.2Ghz 6C 128GB DDR3 2x960GB SSD 1Gbps 100TB - ukusuka $99! Funda malunga
umthombo: www.habr.com