UDavid O'Brien usanda kwethula eyakhe inkampani, i-Xirus (https://xirus.com.au), egxile emikhiqizweni yamafu ye-Microsoft Azure Stack. Zenzelwe ukwakha nokusebenzisa izinhlelo ezixubile ezikhungweni zedatha, izindawo ezisemaphethelweni, amahhovisi akude, kanye namafu.
UDavid uqeqesha abantu nezinkampani kuzo zonke izinto iMicrosoft Azure ne-Azure DevOps (eyayiyi-VSTS) futhi usasebenza ngokubonisana kanye ne-infracoding. Ube ngumnqobi we-Microsoft MVP (Microsoft Most Valuable Professional) Award iminyaka engu-5 futhi muva nje uthole i-Azure MVP Award. Njengomhleli obambisene we-Melbourne Microsoft Cloud kanye ne-Datacentre Meetup, u-O'Brien ukhuluma njalo ezingqungqutheleni zamazwe ngamazwe, ehlanganisa intshisekelo yakhe yokujikeleza umhlaba nothando lokwabelana ngezindaba ze-IT nomphakathi. David's blog itholakala , uphinde ashicilele ukuqeqeshwa kwakhe ku-inthanethi ku-Pluralsight.
Inkulumo ikhuluma ngokubaluleka kwamamethrikhi ekuqondeni ukuthi kwenzekani endaweni yangakini nokuthi isicelo sakho sisebenza kanjani. I-Microsoft Azure inendlela enamandla nelula yokubonisa amamethrikhi azo zonke izinhlobo zemithwalo yemisebenzi, futhi inkulumo ichaza ukuthi ungayisebenzisa kanjani yonke.
Ngo-3 ekuseni ngeSonto, ngenkathi ulele, uvele uvuswe umlayezo wombhalo: "uhlelo lokusebenza olubucayi alusaphenduli futhi." Kwenzakalani? Kuphi futhi siyini isizathu “samabhuleki”? Kule nkulumo, uzofunda ngezinsizakalo iMicrosoft Azure ezinikeza amakhasimende ukuthi aqoqe amalogi futhi, ikakhulukazi, amamethrikhi avela emsebenzini wakho wamafu. U-David uzokutshela ukuthi yimaphi amamethrikhi okufanele ube nentshisekelo kuwo uma usebenza kunkundla yefu nokuthi ungafika kanjani kuwo. Uzofunda ngamathuluzi omthombo ovulekile nokwakhiwa kwedeshibhodi, futhi ugcine usunolwazi olwanele ukuze udale awakho amadeshibhodi.
Futhi uma uvuswa futhi ngo-3 ekuseni ngomlayezo wokuthi uhlelo lokusebenza olubucayi luphahlazekile, ungakwazi ngokushesha ukuthola imbangela yalo.
Sawubona, namuhla sizokhuluma ngamamethrikhi. Igama lami ngingu-David O'Brien, ngingumsunguli kanye nomnikazi wenkampani encane yokubonisana yase-Australia, i-Xirus. Ngiyabonga futhi ngokuza lapha uzochitha isikhathi sakho nami. Pho kungani silapha? Ukukhuluma ngamamethrikhi, noma kunalokho, ngizokutshela ngawo, futhi ngaphambi kokwenza noma yiziphi izinto, ake siqale ngethiyori.
Ngizokutshela ukuthi ayini ama-metrics, yini ongayenza ngawo, yini okudingeka uyinake, indlela yokuqoqa futhi unike amandla ukuqoqwa kwamamethrikhi e-Azure, nokuthi kuyini ukubonwa kwamamethrikhi. Ngizokukhombisa ukuthi lezi zinto zibukeka kanjani efwini le-Microsoft nokuthi usebenza kanjani naleli fu.
Ngaphambi kokuthi siqale, ngizobuza ukuboniswa kwezandla kulabo abasebenzisa iMicrosoft Azure. Ubani osebenza nge-AWS? Ngibona abambalwa. Kuthiwani nge-Google? Ifu le-ALI? Indoda eyodwa! Kuhle. Ngakho ayini amamethrikhi? Incazelo esemthethweni ye-US National Institute of Standards and Technology ithi: “Imethrikhi iyindinganiso yokulinganisa echaza izimo nemithetho yokukala indawo futhi esebenzela ukuqonda imiphumela yokulinganisa.” Kusho ukuthini?
Ake sithathe isibonelo semethrikhi yokushintsha isikhala sediski samahhala somshini obonakalayo. Isibonelo, sinikezwa inombolo engu-90, futhi le nombolo isho amaphesenti, okungukuthi, inani lesikhala samahhala sediski ngu-90%. Ngiyaqaphela ukuthi akuthakazelisi kakhulu ukufunda incazelo yencazelo yamamethrikhi, okuthatha amakhasi angama-40 ngefomethi ye-pdf.
Nokho, imethrikhi ayisho ukuthi umphumela wokulinganisa utholwe kanjani, ibonisa lo mphumela kuphela. Senzani ngamamethrikhi?
Okokuqala, sikala inani lento ukuze sisebenzise umphumela wokulinganisa.
Isibonelo, sithole inani lesikhala samahhala sediski futhi manje singasisebenzisa, sebenzisa le nkumbulo, njll. Uma sesiwutholile umphumela wemethrikhi, kufanele siwuhumushe. Isibonelo, imethrikhi ibuyise umphumela we-90. Sidinga ukwazi ukuthi le nombolo isho ukuthini: inani lesikhala samahhala noma inani lendawo yediski esetshenzisiwe ngamaphesenti noma amagigabhayithi, ukubambezeleka kwenethiwekhi okulingana no-90 ms, njalonjalo, lokho , sidinga ukuhumusha incazelo yenani lemethrikhi. Ukuze amamethrikhi abe nenjongo nhlobo, ngemva kokuhumusha inani elilodwa lemethrikhi, sidinga ukuqinisekisa ukuthi amanani amaningi ayaqoqwa. Lokhu kubaluleke kakhulu ngoba abantu abaningi abasazi isidingo sokuqoqa amamethrikhi. I-Microsoft ikwenze kwaba lula kakhulu ukuqoqa amamethrikhi, kodwa kukuwe ukuthi uqiniseke ukuthi aqoqiwe. Lawa mamethrikhi agcinwa izinsuku ezingama-41 kuphela futhi anyamalale ngosuku lwama-42. Ngakho-ke, kuye ngezakhiwo zemishini yakho yangaphandle noma yangaphakathi, kufanele unakekele indlela yokulondoloza amamethrikhi izinsuku ezingaphezu kwezingu-41 - ngesimo samalogi, izingodo, njll. Ngakho, ngemva kokuqoqwa, kufanele uwabeke endaweni ethile ekuvumela ukuthi ukhiphe zonke izibalo zezinguquko emiphumeleni yemethrikhi uma kudingeka. Uma usuwabeke lapho, ungaqala ukusebenza nawo ngempumelelo.
Kuphela ngemva kokuthi uthole amamethrikhi, uwahumushe futhi uwaqoqe, ungakwazi ukudala isivumelwano seleveli yesevisi ye-SLA. Le SLA kungenzeka ingabi balulekile kangako kumakhasimende akho; kubaluleke kakhulu kosebenza nabo, abaphathi, labo abagcina uhlelo futhi bakhathazekile ngokusebenza kwalo. I-metric ingakala inani lamathikithi - isibonelo, uthola amathikithi angu-5 ngosuku, futhi kulokhu kubonisa isivinini sokuphendula izicelo zabasebenzisi kanye nesivinini sokuxazulula inkinga. Imethrikhi akufanele ivele ithi isayithi lakho lilayisha ngo-20ms noma isivinini sakho sokuphendula singu-20ms, imethrikhi ingaphezu kwenkomba yobuchwepheshe eyodwa.
Ngakho-ke, umsebenzi wengxoxo yethu ukukuletha ngesithombe esinemininingwane yokubaluleka kwamamethrikhi. Imethrikhi isebenza ukuze ngokuyibheka uthole isithombe esiphelele senqubo.
Uma sesinayo imethrikhi, singaqinisekisa ngo-99% ukuthi isistimu iyasebenza, ngoba ayibheki nje ifayela lokungena elithi isistimu iyasebenza. Isiqinisekiso sesikhathi sokuphumula esingu-99% sisho ukuthi, isibonelo, u-99% wesikhathi i-API iphendula ngesivinini esivamile esingu-30 ms. Yilokhu kanye okuthakaselayo abasebenzisi bakho, ozakwenu nabaphathi. Amakhasimende ethu amaningi aqapha izingodo zeseva yewebhu, kodwa awaboni noma yimaphi amaphutha kuwo futhi acabanga ukuthi konke kuhamba kahle. Isibonelo, babona isivinini senethiwekhi esingu-200 Mb/s bese becabanga: “kulungile, konke kuhle!” Kodwa ukufeza lezi 200, abasebenzisi badinga isivinini sokuphendula sama-millisecond angu-30, futhi lokhu kuyinkomba elinganiselwe futhi engaqoqwe kumafayela okungena. Ngesikhathi esifanayo, abasebenzisi bayamangala ukuthi isayithi lilayisha kancane kakhulu, ngoba, lingenawo amamethrikhi adingekayo, abazazi izizathu zalokhu kuziphatha.
Kodwa njengoba sine-100% uptime SLA, amakhasimende aqala ukukhononda ngoba isayithi empeleni linzima kakhulu ukulisebenzisa. Ngakho-ke, ukuze wakhe i-SLA yenhloso, kuyadingeka ukubona isithombe esiphelele senqubo edalwe amamethrikhi aqoqiwe. Lena inkinga eqhubekayo enginayo nabahlinzeki abathile okuthi, lapho benza ama-SLA, bangazi ukuthi igama elithi “uptime” lisho ukuthini futhi ezimweni eziningi abachazi kumakhasimende abo ukuthi i-API yabo isebenza kanjani.
Uma udale isevisi, isibonelo, i-API yomuntu wesithathu, kufanele uqonde ukuthi i-metric ewumphumela ye-39,5 isho ukuthini - impendulo, impendulo ephumelelayo, impendulo ngesivinini esingu-20 ms noma ngesivinini esingu-5 ms. Kukuwe ukuthi uvumelanise i-SLA yabo ne-SLA yakho, kumamethrikhi akho.
Uma usucabangile konke lokhu, ungaqala ukudala ideshibhodi emangalisayo. Ngitshele, ingabe ukhona oseke wasebenzisa uhlelo lokusebenza lwe-Grafana lokubonisa ngeso lengqondo? Kuhle! Ngingumlandeli omkhulu walo mthombo ovulekile ngoba le nto imahhala futhi kulula ukuyisebenzisa.
Uma ungakayisebenzisi i-Grafana okwamanje, ngizokutshela ukuthi usebenza kanjani nayo. Noma ubani owazalwa ngeminyaka yama-80s nama-90s mhlawumbe ukhumbula i-CareBears? Angazi ukuthi la mabhere ayedume kangakanani eRussia, kodwa uma kuziwa kumamethrikhi, kufanele sibe “amabhere anakekelayo” afanayo. Njengoba ngishilo, udinga isithombe esikhulu sokuthi lonke uhlelo lusebenza kanjani, futhi akufanele nje kube nge-API yakho, iwebhusayithi yakho, noma isevisi esebenza emshinini obonakalayo.
Kufanele uhlele ukuqoqwa kwalawo mamethrikhi abonisa ngokugcwele ukusebenza kwayo yonke isistimu. Iningi lenu lingabathuthukisi be-software, ngakho-ke impilo yenu ishintsha njalo, ijwayelana nezimfuneko zomkhiqizo omusha, futhi njengoba nje ukhathazekile ngezinqubo zokubhala amakhodi, kufanele ukhathazeke ngamamethrikhi. Udinga ukwazi ukuthi imethrikhi ihlobana kanjani nawo wonke umugqa wekhodi oyibhalayo. Isibonelo, ngesonto elizayo uqala umkhankaso omusha wokumaketha futhi ulindele inani elikhulu labasebenzisi ukuthi bavakashele isayithi lakho. Ukuze uhlaziye lo mcimbi, uzodinga amamethrikhi, futhi ungase udinge yonke ideshibhodi ukuze ulandelele umsebenzi walaba bantu. Uzodinga amamethrikhi ukuze uqonde ukuthi uphumelele kangakanani umkhankaso wakho wokumaketha nokuthi usebenza kanjani ngempela. Bazokusiza, isibonelo, ukuthuthukisa i-CRM ephumelelayo - uhlelo lokuphatha ubudlelwano bekhasimende.
Ngakho-ke ake siqale ngenkonzo yethu yefu ye-Azure. Kulula kakhulu ukuthola nokuhlela ukuqoqwa kwamamethrikhi ngoba ine-Azure Monitor. Leli qapha libeka phakathi ukuphathwa kokucushwa kwesistimu yakho. Isici ngasinye se-Azure ofuna ukusisebenzisa ohlelweni lwakho sinamamethrikhi amaningi anikwe amandla ngokuzenzakalela. Lolu uhlelo lokusebenza lwamahhala olusebenza ngaphandle kwebhokisi futhi aludingi izilungiselelo zokuqala; awudingi ukubhala noma "ukukulungela" noma yini ohlelweni lwakho. Sizoqinisekisa lokhu ngokubheka idemo elandelayo.
Ngaphezu kwalokho, kungenzeka ukuthumela lawa mamethrikhi kuzinhlelo zokusebenza zezinkampani zangaphandle, ezifana ne-Splunk log storage kanye nesistimu yokuhlaziya, uhlelo lokusebenza lokuphathwa kwelogi olusekelwe efwini i-SumoLogic, ithuluzi lokucubungula ilogi ye-ELK, ne-IBM Radar. Yiqiniso, kunomehluko omncane oncike kuzinsiza ozisebenzisayo - umshini obonakalayo, izinsizakalo zenethiwekhi, i-Azure SQL yolwazi, okungukuthi, ukusetshenziswa kwamamethrikhi kuyahluka kuye ngemisebenzi yendawo yakho yokusebenza. Ngeke ngisho ukuthi lezi zingxabano zibucayi, kodwa, ngeshwa, zisekhona, futhi lokhu kufanele kucatshangelwe. Ukunika amandla nokuthumela amamethrikhi kungenzeka ngezindlela ezimbalwa: nge-Portal, CLI/Power Shell, noma kusetshenziswa izifanekiso ze-ARM.
Ngaphambi kokuthi ngiqale idemo yami yokuqala, ngizophendula noma yimiphi imibuzo ongase ube nayo. Uma ingekho imibuzo, ake siqale. Isikrini sibonisa ukuthi ikhasi le-Azure Monitor libukeka kanjani. Ukhona yini kini othi leli qapha alisebenzi?
Ngakho-ke manje konke kuhamba kahle, ungabona ukuthi izinsizakalo zokuqapha zibukeka kanjani. Ngingasho ukuthi leli yithuluzi elihle kakhulu futhi elilula kakhulu lomsebenzi wansuku zonke. Ingasetshenziselwa ukuqapha izinhlelo zokusebenza, amanethiwekhi nengqalasizinda. Muva nje, isixhumi esibonakalayo sokuqapha sithuthukisiwe, futhi uma izinsiza zangaphambili bezitholakala ezindaweni ezihlukene, manje lonke ulwazi lwezinsizakalo luhlanganiswa ekhasini lasekhaya lomqaphi.
Ithebula lamamethrikhi liyithebhu esendleleni ye-HomeMonitorMetrics, ongaya kuyo ukuze ubone wonke ama-metrics atholakalayo bese ukhetha lawo owadingayo. Kodwa uma udinga ukunika amandla ukuqoqwa kwamamethrikhi, udinga ukusebenzisa indlela yohla lwemibhalo yezilungiselelo ze-HomeMonitorDiagnostic futhi uhlole amabhokisi okuhlola amamethrikhi Avunyelwe/Akhubazekile. Ngokuzenzakalela, cishe wonke amamethrikhi anikwe amandla, kodwa uma udinga ukunika amandla okuthile okwengeziwe, uzodinga ukushintsha isimo sokuxilonga sisuke kokuthi Kukhubazekile siye kokuthi Kunikwe amandla.
Ukuze wenze lokhu, chofoza umugqa wemethrikhi ekhethiwe nakuthebhu evulayo, nika amandla imodi yokuxilonga. Uma uzohlaziya imethrikhi ekhethiwe, ngemva kokuchofoza okuthi Vula isixhumanisi sokuxilonga, udinga ukuhlola ibhokisi lokuhlola elithi Thumela ku-Log Analytics efasiteleni elivelayo.
I-Log Analytics ifana kancane ne-Splunk, kodwa ibiza kancane. Le sevisi ikuvumela ukuthi uqoqe wonke amamethrikhi akho, amalogu nakho konke okunye okudingayo futhi ukubeke endaweni yokusebenza ye-Log Analytics. Isevisi isebenzisa ulimi olukhethekile lokucubungula imibuzo ye-KQL - Ulimi lwe-Kusto Quarry, sizobheka umsebenzi wayo kudemo elandelayo. Okwamanje, ngizoqaphela ukuthi ngosizo lwayo ungakwazi ukudala imibuzo mayelana namamethrikhi, amalogi, imigomo, izitayela, amaphethini, njll. futhi udale amadeshibhodi.
Ngakho-ke, sihlola ibhokisi elithi Thumela ku-Log Analytics kanye namabhokisi okuhlola ephaneli YELOG: DataPlaneRequests, MongoRequests kanye ne-QueryRuntimeStatistics, futhi ngezansi kuphaneli ye-METRIC - ibhokisi lokuhlola elithi Izicelo. Bese sinikeza igama bese silondoloza izilungiselelo. Emugqeni womyalo, lokhu kumelela imigqa emibili yekhodi. Ngendlela, igobolondo le-Azure Cloud ngalo mqondo lifana ne-Google, elikuvumela futhi ukuthi usebenzise umugqa womyalo kusiphequluli sakho sewebhu. I-AWS ayinayo into enjalo, ngakho-ke i-Azure ilula kakhulu ngalo mqondo.
Isibonelo, ngingasebenzisa idemo ngesixhumi esibonakalayo sewebhu ngaphandle kokusebenzisa noma iyiphi ikhodi kukhompuyutha yami ephathekayo. Ukwenza lokhu, kufanele ngiqinisekise nge-akhawunti yami ye-Azure. Khona-ke ungasebenzisa, isibonelo, i-terrafone, uma usuvele uyisebenzisa, linda uxhumano lwesevisi futhi uthole indawo yokusebenza ye-Linux esetshenziswa yi-Microsoft ngokuzenzakalelayo.
Okulandelayo, ngisebenzisa i-Bash, eyakhelwe ku-Azure Cloud Shell. Into ewusizo kakhulu i-IDE eyakhelwe kusiphequluli, inguqulo engasindi ye-VS Code. Okulandelayo, ngingangena kusifanekiso sami semethrikhi yephutha, ngisihlele, futhi ngisenze ngokwezifiso ukuze sihambisane nezidingo zami.
Uma ususethe iqoqo lamamethrikhi kulesi sifanekiso, ungasisebenzisa ukuze udale amamethrikhi ayo yonke ingqalasizinda yakho. Uma sesisebenzise amamethrikhi, siwaqoqile, futhi sawagcina, kuzodingeka siwabone ngeso lengqondo.
I-Azure Monitor isebenza kuphela ngamamethrikhi futhi ayinikezi isithombe esiphelele sempilo yohlelo lwakho. Ungase ube nenani lezinye izinhlelo zokusebenza ezisebenza ngaphandle kwemvelo ye-Azure. Ngakho-ke uma udinga ukuqapha zonke izinqubo, ukubona ngeso lengqondo wonke amamethrikhi aqoqwe endaweni eyodwa, khona-ke i-Azure Monitor ayikulungele lokhu.
Ukuze kuxazululwe le nkinga, i-Microsoft inikeza ithuluzi le-Power BI, isofthiwe ephelele yokuhlaziywa kwebhizinisi ehlanganisa ukubonwa kwedatha enhlobonhlobo. Lona umkhiqizo obiza kakhulu, izindleko zawo zincike kusethi yemisebenzi oyidingayo. Ngokuzenzakalelayo, ikunikeza izinhlobo ezingama-48 zedatha okufanele uyicubungule futhi ixhunywe ku-Azure SQL Data Warehouses, I-Azure Data Lake Storage, Izinsiza Zokufunda Zomshini we-Azure, kanye ne-Azure Databricks. Ngokusebenzisa i-scalability, ungathola idatha entsha njalo ngemizuzu engama-30. Lokhu kungase kunganele noma kunganele izidingo zakho uma udinga ukubonwa kokuqapha ngesikhathi sangempela. Kulokhu, kunconywa ukusebenzisa izinhlelo zokusebenza ezifana neGrafana engiyishilo. Ngaphezu kwalokho, imibhalo ye-Microsoft ichaza ikhono lokuthumela amamethrikhi, amalogi namatafula omcimbi kusetshenziswa amathuluzi e-SIEM ezinhlelweni zokubuka ngeso I-Splunk, i-SumoLogic, i-ELK ne-IBM radar.
23:40 imiz
Izoqhubeka maduze...
Ezinye izikhangiso 🙂
Siyabonga ngokuhlala nathi. Uyazithanda izindatshana zethu? Ufuna ukubona okuqukethwe okuthakaselayo okwengeziwe? Sisekele ngokufaka i-oda noma ngokuncoma kubangani, , i-analogue ehlukile yamaseva ezinga lokungena, esungulwe yithi ngenxa yakho: (itholakala nge-RAID1 kanye ne-RAID10, kufika kuma-cores angu-24 kuze kufike ku-40GB DDR4).
I-Dell R730xd 2x ishibhile esikhungweni sedatha se-Equinix Tier IV e-Amsterdam? Lapha kuphela eNetherlands! I-Dell R420 - 2x E5-2430 2.2Ghz 6C 128GB DDR3 2x960GB SSD 1Gbps 100TB - isuka ku-$99! Funda mayelana
Source: www.habr.com