Ua hoʻokumu ʻo David O'Brien i kāna hui ponoʻī, ʻo Xirus (https://xirus.com.au), e kālele ana i nā huahana ao Microsoft Azure Stack. Hoʻolālā ʻia lākou e kūkulu mau a holo i nā noi hybrid ma nā kikowaena ʻikepili, nā wahi lihi, nā keʻena mamao, a me ke ao.
Hoʻomaʻamaʻa ʻo David i nā poʻe a me nā ʻoihana ma nā mea āpau Microsoft Azure a me Azure DevOps (VSTS ma mua) a ke hana mau nei i nā kūkākūkā lima a me ka infracoding. Ua lanakila ʻo ia ma Microsoft MVP (Microsoft Most Valuable Professional) Award no 5 mau makahiki a ua loaʻa hou iā ia ka Azure MVP Award. Ma ke ʻano he hui hui o ka Melbourne Microsoft Cloud a me Datacentre Meetup, ʻōlelo mau ʻo O'Brien i nā ʻaha kūkā honua, e hui pū ana i kona hoihoi i ka huakaʻi i ka honua me ka makemake e kaʻana like i nā moʻolelo IT me ke kaiāulu. Aia ka moʻomanaʻo a David ma , hoʻopuka pū ʻo ia i kāna aʻo pūnaewele ma Pluralsight.
Kūkākūkā ke kamaʻilio e pili ana i ke koʻikoʻi o nā metric i ka hoʻomaopopo ʻana i nā mea e hana nei i kou kaiapuni a me ke ʻano o kāu noi. He ala ikaika a maʻalahi ʻo Microsoft Azure e hōʻike i nā metric no nā ʻano hana like ʻole, a wehewehe ka haʻiʻōlelo pehea e hiki ai iā ʻoe ke hoʻohana iā lākou āpau.
Ma ka hola 3 a.m. o ka lā Sābati, ʻoiai ʻoe e hiamoe ana, ua hoʻāla koke ʻia ʻoe e kahi leka uila: "ʻaʻole pane hou ka app supercritical." He aha ana lā? Ma hea a he aha ke kumu o ka "brakes"? Ma kēia kamaʻilio ʻana, e aʻo ʻoe e pili ana i nā lawelawe a Microsoft Azure e hāʻawi aku ai i nā mea kūʻai aku e hōʻiliʻili i nā lāʻau a, ʻo ia hoʻi, nā metric mai kāu mau hana kapuaʻi. E haʻi aku ʻo David iā ʻoe i nā metric āu e makemake ai i ka wā e hana ai ma kahi kahua ao a pehea e hiki ai iā lākou. E aʻo ʻoe e pili ana i nā mea hana open source a me ke kūkulu ʻana i ka dashboard, a hoʻopau me ka ʻike lawa e hana ai i kāu dashboards ponoʻī.
A inā hoʻāla hou ʻia ʻoe i ka hola 3 a.m. e kahi leka i hāʻule kahi noi koʻikoʻi, hiki iā ʻoe ke ʻike koke i kona kumu.
Aloha ahiahi, i kēia lā e kamaʻilio mākou e pili ana i nā metric. ʻO David O'Brien koʻu inoa, ʻo wau ka mea hoʻokumu a me ka mea nona kahi hui kūkākūkā liʻiliʻi o Australia, ʻo Xirus. Mahalo hou no kou hele ʻana mai e hoʻolilo i kou manawa me aʻu. No laila, no ke aha mākou i ʻaneʻi? No ke kamaʻilio e pili ana i nā metric, a i ʻole, e haʻi wau iā ʻoe e pili ana iā lākou, a ma mua o ka hana ʻana i kekahi mea, e hoʻomaka kākou me ka manaʻo.
E haʻi aku wau iā ʻoe he aha nā metric, he aha kāu e hana ai me lākou, ka mea e pono ai ʻoe e hoʻolohe, pehea e hōʻiliʻili ai a hiki ai i ka hōʻiliʻili metric ma Azure, a he aha ka metric visualization. E hōʻike aku wau iā ʻoe i ke ʻano o kēia mau mea ma ke ao Microsoft a pehea e hana ai me kēia ao.
Ma mua o ka hoʻomaka ʻana, e nīnau wau i kahi hōʻike lima mai ka poʻe hoʻohana iā Microsoft Azure. ʻO wai e hana pū me AWS? Ke ike nei au he kakaikahi. Pehea e pili ana iā Google? ALI Kapua? Hoʻokahi kanaka! Nui. No laila he aha nā metric? ʻO ka wehewehe kūhelu o ka US National Institute of Standards and Technology ʻo ia: "ʻO ka metric kahi maʻamau ana e wehewehe ana i nā kūlana a me nā lula no ke ana ʻana i kahi waiwai a e hoʻomaopopo i nā hopena ana." He aha ka manaʻo?
E lawe kākou i kekahi laʻana o kahi metric no ka hoʻololi ʻana i ka wahi diski manuahi o kahi mīkini virtual. No ka laʻana, hāʻawi ʻia mākou i ka helu 90, a ʻo kēia helu ke ʻano o ka pākēneka, ʻo ia hoʻi, ʻo ka nui o ka wahi disk manuahi he 90%. Hoʻomaopopo wau ʻaʻole hoihoi loa ka heluhelu ʻana i ka wehewehe ʻana i ka wehewehe ʻana o nā metric, e lawe ana i nā ʻaoʻao 40 i ka format pdf.
Eia naʻe, ʻaʻole ʻōlelo ka metric pehea i loaʻa ai ka hopena ana, hōʻike wale ia i kēia hopena. He aha kā mākou e hana ai me nā metric?
ʻO ka mua, ana mākou i ka waiwai o kekahi mea i mea e hoʻohana ai i ka hopena ana.
No ka laʻana, ua ʻike mākou i ka nui o ka hakahaka manuahi a hiki iā mākou ke hoʻohana iā ia, hoʻohana i kēia hoʻomanaʻo, etc. Ke loaʻa iā mākou ka hopena metric, pono mākou e wehewehe. No ka laʻana, ua hoʻihoʻi ka metric i ka hopena o 90. Pono mākou e ʻike i ke ʻano o kēia helu: ka nui o ka hakahaka kaʻawale a i ʻole ka nui o ka wahi diski i hoʻohana ʻia ma ka pakeneka a i ʻole gigabytes, ka latency pūnaewele e like me 90 ms, a pēlā aku, ʻo ia. , pono mākou e wehewehe i ke ʻano o ka waiwai metric. I mea e pono ai nā metric, ma hope o ka unuhi ʻana i kahi waiwai metric hoʻokahi, pono mākou e hōʻoia e hōʻiliʻili ʻia nā waiwai he nui. He mea nui kēia no ka mea ʻaʻole ʻike ka poʻe he nui i ka pono e hōʻiliʻili i nā metric. Ua maʻalahi loa ʻo Microsoft e hōʻiliʻili i nā metric, akā aia iā ʻoe e hōʻoia e ʻohi ʻia lākou. Mālama ʻia kēia mau ana no nā lā 41 wale nō a nalowale i ka lā 42. No laila, e pili ana i nā waiwai o kāu lako waho a i loko paha, pono ʻoe e mālama i ke ʻano o ka mālama ʻana i nā metric no nā lā ʻoi aku ma mua o 41 mau lā - ma ke ʻano o nā logs, logs, etc. No laila, ma hope o ka hōʻiliʻili ʻana, pono ʻoe e kau iā lākou ma kekahi wahi e hiki ai iā ʻoe ke huki i nā helu āpau o nā loli i nā hopena metric inā pono. Ke kau ʻoe iā lākou ma laila, hiki iā ʻoe ke hoʻomaka e hana pū me lākou me ka maikaʻi.
Ma hope wale nō o ka loaʻa ʻana o nā ana, wehewehe a hōʻiliʻili iā lākou, hiki iā ʻoe ke hana i kahi ʻaelike SLA - lawelawe pae. ʻAʻole paha he mea koʻikoʻi kēia SLA i kāu mea kūʻai aku; ʻoi aku ka mea nui i kāu mau hoa hana, nā luna, nā mea mālama i ka ʻōnaehana a hopohopo e pili ana i kāna hana. Hiki i ka metric ke ana i ka helu o nā tiketi - no ka laʻana, loaʻa iā ʻoe nā tiketi 5 i kēlā me kēia lā, a ma kēia hihia e hōʻike ana i ka wikiwiki o ka pane ʻana i nā noi mea hoʻohana a me ka wikiwiki o ka hoʻoponopono pilikia. ʻAʻole pono e ʻōlelo wale ke metric e hoʻouka ana kāu pūnaewele i 20ms a i ʻole 20ms ka wikiwiki o kāu pane, ʻoi aku ka metric ma mua o hoʻokahi hōʻailona ʻenehana.
No laila, ʻo ka hana o kā mākou kamaʻilio ʻana e hōʻike iā ʻoe i kahi kiʻi kikoʻī o ke ʻano o nā metric. Hoʻohana ka metric i hiki iā ʻoe ke kiʻi i ke kiʻi piha o ke kaʻina hana ma ka nānā ʻana iā ia.
Ke loaʻa iā mākou ka metric, hiki iā mākou ke hōʻoia 99% e hana ana ka ʻōnaehana, no ka mea, ʻaʻole ia e nānā wale i kahi faila log e ʻōlelo ana e hana ana ka ʻōnaehana. ʻO kahi hōʻoia 99% uptime ʻo ia hoʻi, no ka laʻana, 99% o ka manawa e pane ai ka API i ka wikiwiki maʻamau o 30 ms. ʻO kēia ka mea hoihoi i kāu mea hoʻohana, kāu mau hoa hana a me nā luna. Ka nui o kā mākou mea kūʻai e nānā i nā log server pūnaewele, akā ʻaʻole lākou i ʻike i kekahi hewa i loko o lākou a manaʻo lākou ua maikaʻi nā mea a pau. No ka laʻana, ʻike lākou i ka wikiwiki o ka pūnaewele o 200 Mb/s a noʻonoʻo: "ʻAe, maikaʻi nā mea āpau!" Akā no ka hoʻokō ʻana i kēia mau 200, pono nā mea hoʻohana i ka wikiwiki pane o 30 milliseconds, a ʻo kēia ka mea hōʻailona ʻaʻole i ana ʻia a ʻaʻole i hōʻiliʻili ʻia i nā faila log. I ka manawa like, kahaha nā mea hoʻohana i ka lohi ʻana o ka pūnaewele, no ka mea, ʻaʻole i loaʻa nā metric pono, ʻaʻole lākou ʻike i nā kumu o kēia ʻano.
Akā no ka loaʻa ʻana o ka 100% uptime SLA, hoʻomaka nā mea kūʻai aku e hoʻopiʻi no ka mea paʻakikī loa ka pūnaewele e hoʻohana. No laila, no ka hana ʻana i kahi SLA pahuhopu, pono e ʻike i ke kiʻi piha o ke kaʻina hana i hana ʻia e nā metric i hōʻiliʻili ʻia. He pilikia hoʻomau kēia i loaʻa iaʻu me kekahi mau mea hoʻolako, i ka wā e hana ai i nā SLA, ʻaʻole maopopo i ke ʻano o ka huaʻōlelo "uptime" a ma ka hapa nui o nā hihia ʻaʻole wehewehe i kā lākou mea kūʻai aku pehea e hana ai kā lākou API.
Inā hana ʻoe i kahi lawelawe, no ka laʻana, he API no ke kolu o ke kanaka, pono ʻoe e hoʻomaopopo i ke ʻano o ka metric hopena o 39,5 - pane, pane kūleʻa, pane i ka wikiwiki 20 ms a i ʻole 5 ms ka wikiwiki. Aia iā ʻoe ke hoʻololi i kā lākou SLA i kāu SLA ponoʻī, i kāu metric ponoʻī.
Ke hoʻomaopopo ʻoe i kēia mau mea a pau, hiki iā ʻoe ke hoʻomaka i ka hana ʻana i kahi dashboard kupanaha. E haʻi mai iaʻu, ua hoʻohana mua paha kekahi i ka palapala noiʻi ʻike kiʻi kiʻi ʻo Grafana? Nui! Makemake nui au i kēia puna hāmama no ka mea manuahi a maʻalahi kēia mea.
Inā ʻaʻole ʻoe i hoʻohana iā Grafana, e haʻi wau iā ʻoe pehea e hana ai me ia. Hoʻomanaʻo paha ka poʻe i hānau i ka makahiki 80 a me 90 iā CareBears? ʻAʻole au i ʻike i ka kaulana o kēia mau bea ma Rūsia, akā i ka wā e pili ana i nā metric, pono mākou e like me nā "bea mālama." E like me kaʻu i ʻōlelo ai, pono ʻoe i kahi kiʻi nui o ka hana ʻana o ka ʻōnaehana holoʻokoʻa, ʻaʻole pono ia e pili ana i kāu API, kāu pūnaewele, a i ʻole ka lawelawe e holo ana i kahi mīkini virtual.
Pono ʻoe e hoʻonohonoho i ka hōʻiliʻili ʻana o kēlā mau ana i hōʻike piha loa i ka hana o ka ʻōnaehana holoʻokoʻa. ʻO ka hapa nui o ʻoukou he mea hoʻomohala polokalamu, no laila ke loli mau nei kou ola, e hoʻololi i nā koi huahana hou, a e like me kou hopohopo i nā kaʻina coding, pono ʻoe e hopohopo i nā metric. Pono ʻoe e ʻike pehea e pili ai ka metric i kēlā me kēia laina code āu e kākau ai. No ka laʻana, i kēia pule aʻe e hoʻomaka ana ʻoe i kahi hoʻolaha kūʻai hou a manaʻo i ka nui o nā mea hoʻohana e kipa i kāu pūnaewele. No ka nānā ʻana i kēia hanana, pono ʻoe i nā metric, a pono paha ʻoe i kahi dashboard holoʻokoʻa e nānā i ka hana a kēia poʻe. Pono ʻoe i nā metric e hoʻomaopopo i ka kūleʻa o kāu hoʻolaha kūʻai aku a pehea e hana maoli ai. E kōkua lākou iā ʻoe, no ka laʻana, e hoʻomohala i kahi CRM maikaʻi - ʻōnaehana hoʻokele pili pili.
No laila e hoʻomaka kākou me kā mākou lawelawe kapuaʻi Azure. He mea maʻalahi loa ka loaʻa a hoʻonohonoho ʻana i ka hōʻiliʻili metric no ka mea he Azure Monitor. Hoʻonohonoho kēia mea nānā i kāu hoʻonohonoho hoʻonohonoho hoʻonohonoho. ʻO kēlā me kēia o nā mea Azure āu e makemake ai e hoʻopili i kāu ʻōnaehana he nui nā metric i hoʻohana ʻia e ka paʻamau. He noi manuahi kēia e hana pololei ana mai loko mai o ka pahu a ʻaʻole koi i nā hoʻonohonoho mua; ʻaʻole pono ʻoe e kākau a i ʻole "screw" i kekahi mea i kāu ʻōnaehana. E hōʻoia mākou i kēia ma ka nānā ʻana i ka demo aʻe.
Eia hou, hiki ke hoʻouna aku i kēia mau ana i nā noi ʻaoʻao ʻekolu, e like me ka Splunk log storage and analysis system, ka palapala hoʻokele log log-based SumoLogic, ka mea hana hana log ELK, a me IBM Radar. ʻOiaʻiʻo, aia nā ʻokoʻa liʻiliʻi e hilinaʻi ana i nā kumuwaiwai āu e hoʻohana ai - kahi mīkini virtual, nā lawelawe pūnaewele, nā ʻikepili Azure SQL, ʻo ia hoʻi, ʻokoʻa ka hoʻohana ʻana i nā metric ma muli o nā hana o kāu wahi hana. ʻAʻole wau e ʻōlelo he koʻikoʻi kēia mau ʻokoʻa, akā, ʻo ka mea pōʻino, aia nō lākou, a pono e noʻonoʻo ʻia kēia. Hiki ke hoʻohana a hoʻouna i nā ana ma nā ʻano he nui: ma o Portal, CLI/Power Shell, a i ʻole ka hoʻohana ʻana i nā template ARM.
Ma mua o koʻu hoʻomaka ʻana i kaʻu demo mua, e pane wau i nā nīnau āu e loaʻa ai. Inā ʻaʻohe nīnau, e hoʻomaka kākou. Hōʻike ka pale i ke ʻano o ka ʻaoʻao Azure Monitor. Hiki i kekahi o ʻoukou ke ʻōlelo ʻaʻole hana kēia monitor?
No laila ua maikaʻi nā mea a pau, hiki iā ʻoe ke ʻike i ke ʻano o nā lawelawe monitor. Hiki iaʻu ke ʻōlelo he mea hana maikaʻi a maʻalahi loa kēia no ka hana o kēlā me kēia lā. Hiki ke hoʻohana ʻia e nānā i nā noi, nā pūnaewele a me nā ʻōnaehana. I kēia mau lā, ua hoʻomaikaʻi ʻia ka nānā ʻana, a inā i loaʻa nā lawelawe ma nā wahi like ʻole, i kēia manawa ua hoʻohui ʻia nā ʻike āpau e pili ana i nā lawelawe ma ka ʻaoʻao home o ka monitor.
ʻO ka papa metrics he tab ma ke ala HomeMonitorMetrics, kahi āu e hele ai e ʻike i nā metric āpau i loaʻa a koho i nā mea āu e pono ai. Akā inā pono ʻoe e ʻae i ka hōʻiliʻili metric, pono ʻoe e hoʻohana i ke ala papa kuhikuhi hoʻonohonoho HomeMonitorDiagnostic a nānā i nā pahu kaha Enabled/Disabled. Ma ka maʻamau, aneane hiki ke hoʻohana ʻia nā metric āpau, akā inā pono ʻoe e hoʻā i kahi mea hou aku, pono ʻoe e hoʻololi i ke kūlana diagnostic mai Disabled i Enabled.
No ka hana ʻana i kēia, kaomi ma ka laina o ka metric i koho ʻia a ma ka ʻaoʻao e wehe ana, hiki i ke ʻano diagnostic. Inā ʻoe e nānā i ka metric i koho ʻia, a laila ma hope o ke kaomi ʻana i ka loulou Turn on diagnostic, pono ʻoe e nānā i ka pahu hoʻouna i ka Log Analytics ma ka puka aniani e kū mai ana.
Ua like like ka Log Analytics me Splunk, akā emi nā kumukūʻai. Hāʻawi kēia lawelawe iā ʻoe e hōʻiliʻili i kāu mau metric āpau, nā lāʻau a me nā mea ʻē aʻe āu e pono ai a waiho iā lākou i ka papa hana Log Analytics. Hoʻohana ka lawelawe i kahi ʻōlelo hoʻoponopono nīnau KQL kūikawā - Kusto Quarry Language, e nānā mākou i kāna hana ma ka demo aʻe. I kēia manawa, e hoʻomaopopo wau me kāna kōkua hiki iā ʻoe ke hana i nā nīnau e pili ana i nā metrics, logs, terms, trends, patterns, etc. a hana i nā dashboards.
No laila, nānā mākou i ka pahu Hoʻouna iā Log Analytics a me nā pahu pahu LOG panel: DataPlaneRequests, MongoRequests a me QueryRuntimeStatistics, a ma lalo o ka METRIC panel - ka pahu noi noi. A laila hāʻawi mākou i kahi inoa a mālama i nā hoʻonohonoho. Ma ka laina kauoha, hōʻike kēia i ʻelua laina code. Ma ke ala, ʻo ka shell Azure Cloud i kēia ʻano like me Google, kahi e hiki ai iā ʻoe ke hoʻohana i ka laina kauoha i kāu polokalamu kele pūnaewele. ʻAʻohe mea like me AWS, no laila ʻoi aku ka maʻalahi o Azure ma kēia ʻano.
No ka laʻana, hiki iaʻu ke holo i kahi demo ma o ka pūnaewele pūnaewele me ka hoʻohana ʻole i kekahi code ma kaʻu kamepiula. No ka hana ʻana i kēia, pono wau e hōʻoia me kaʻu moʻokāki Azure. A laila hiki iā ʻoe ke hoʻohana, no ka laʻana, terrafone, inā ʻoe e hoʻohana mua iā ia, e kali no ka pili ʻana i ka lawelawe a loaʻa i ka ʻenehana hana Linux e hoʻohana ai ʻo Microsoft ma ka paʻamau.
A laila, hoʻohana wau i Bash, i kūkulu ʻia i loko o ka Azure Cloud Shell. ʻO kahi mea maikaʻi loa ka IDE i kūkulu ʻia i loko o ka polokalamu kele pūnaewele, kahi māmā māmā o VS Code. Ma hope aʻe, hiki iaʻu ke komo i loko o kaʻu template metric error, hoʻoponopono iā ia, a hoʻonohonoho pono i kaʻu mau pono.
Ke hoʻonohonoho ʻoe i ka hōʻiliʻili metric i kēia template, hiki iā ʻoe ke hoʻohana iā ia e hana i nā ana no kāu ʻōnaehana holoʻokoʻa. Ke hoʻohana mākou i nā ana, hōʻiliʻili iā lākou, a mālama iā lākou, pono mākou e nānā iā lākou.
Hana wale ʻo Azure Monitor i nā metric a ʻaʻole hāʻawi i kahi kiʻi holoʻokoʻa o ke olakino o kāu ʻōnaehana. Loaʻa paha iā ʻoe kekahi mau noi ʻē aʻe e holo ana ma waho o ke kaiapuni Azure. No laila inā pono ʻoe e nānā i nā kaʻina hana a pau, e nānā i nā metric i hōʻiliʻili ʻia ma kahi hoʻokahi, a laila ʻaʻole kūpono ʻo Azure Monitor no kēia.
No ka hoʻoponopono ʻana i kēia pilikia, hāʻawi ʻo Microsoft i ka hāmeʻa Power BI, kahi lako polokalamu piha no ka nānā ʻana i ka ʻoihana e pili ana i ka ʻike ʻana i nā ʻano ʻikepili ākea. He huahana maikaʻi loa kēia, ʻo ke kumukūʻai e pili ana i ka hoʻonohonoho o nā hana āu e pono ai. Ma ka maʻamau, hāʻawi ia iā ʻoe i nā ʻano ʻikepili 48 e hana ai a pili ʻia i nā Azure SQL Data Warehouses, Azure Data Lake Storage, Azure Machine Learning Services, a me Azure Databricks. Ke hoʻohana nei i ka scalability, hiki iā ʻoe ke loaʻa i nā ʻikepili hou i kēlā me kēia 30 mau minuke. Ua lawa paha kēia no kāu mau pono inā makemake ʻoe i ka nānā ʻana i ka manawa maoli. I kēia hihia, ʻōlelo ʻia e hoʻohana i nā noi e like me ka Grafana aʻu i ʻōlelo ai. Eia kekahi, wehewehe ʻo Microsoft palapala i ka hiki ke hoʻouna i nā metrics, logs a me nā papa hanana me ka hoʻohana ʻana i nā mea hana SIEM i nā ʻōnaehana hiʻohiʻona Splunk, SumoLogic, ELK a me IBM radar.
23:40 min
E hoʻomau koke ʻia ...
Kekahi mau hoʻolaha 🙂
Mahalo no kou noho pū ʻana me mākou. Makemake ʻoe i kā mākou ʻatikala? Makemake ʻoe e ʻike i nā mea hoihoi hou aʻe? E kākoʻo iā mākou ma ke kau ʻana i kahi kauoha a i ʻole ka ʻōlelo ʻana i nā hoaaloha, , he analogue kū hoʻokahi o nā kikowaena hoʻokomo-level, i hana ʻia e mākou no ʻoe: (loaʻa me RAID1 a me RAID10, a hiki i 24 cores a hiki i 40GB DDR4).
ʻO Dell R730xd 2 mau manawa maʻalahi ma Equinix Tier IV kikowaena data ma Amsterdam? Eia wale nō ma Netherlands! Dell R420 - 2x E5-2430 2.2Ghz 6C 128GB DDR3 2x960GB SSD 1Gbps 100TB - mai $99! Heluhelu e pili ana
Source: www.habr.com