Interactive AIOps - Speaker Deck

Interactive AIOps

by Yuuki Tsubouchi (yuuk1)

Slide 1

Slide 1 text

Interactive AIOps ͘͞ΒΠϯλʔωοτݚڀॴ߹॓202 1 @yuuk1t

Slide 2

Slide 2 text

IPSJ-ONE in 2017 ιϑτ΢ΣΞγεςϜ͕ࣗݾ؍ଌͱࣗݾ࣮ݧΛ  ܁Γฦ͠ɺࣗ཯తʹֶश͍ͯ͘͠γεςϜల๬ΛఏҊ

Slide 3

Slide 3 text

ΦϖϨʔλʔ Ϣʔβʔ ITγεςϜ AI SRE ΦϖϨʔλʔ͕Ͳͷఔ౓ϦεΫΛ   ͱΕΔ͔Λઃఆ Chaos Engineering Experimentable Infrastructure

Slide 4

Slide 4 text

ݚڀϏδϣϯΛ΋ͭ • ձࣾͷ՝୊ղܾ͔Βɺݸਓͱͯ͠ͷʮ୳ڀʯ΁ • Ϗδϣϯ͸௕ظతݟ஍ʹͨͬͨߴ͍ந৅౓Λ΋ͭͨΊɺϏδϣ ϯͱؔ࿈͚༷ͮͯʑͳࣄฑʹڵຯΛ΋ͭΑ͏ʹͳͬͨ • ࣗ෼͝ͱԽͰ͖ΔΪϦΪϦͷ۩ମੑΛ΋ͨͤΔ͜ͱͰɺͿΕͣ ʹݚڀΛਐΊΒΕΔ

Slide 5

Slide 5 text

ߏ૝͔Β5೥ • γεςϜΛʮ؍ଌʯ͢ΔͨΊͷݚڀΛਐΊ͖ͯͨ • HeteroTSDB since 201 6 • Transtracer since 201 8 • ʮ࣮ݧʯΛओମͱ͢Δݚڀʹγϑτ • ࣮ݧʹΑΔσʔληοτੜ੒ Meltria since 202 1 • ʮ࣮ݧʯʹΑΓੜ੒ͨ͠σʔλΛ౷ܭɾػցֶश • σʔληοτͷલॲཧ TSifter since 202 0 • ݪҼ਍அ [H. Tsuruta 2021 ] • γεςϜֶ͕शͯ͠ݡ͘ͳ͍ͬͯ͘͜ͱ͸·ͩͰ͖͍ͯͳ͍

Slide 6

Slide 6 text

AIOps Artificial Intelligence for IT Operations (AIOps) investigates the use of Artificial Intelligence (AI) for the management and improvement of IT services Notaro, Paolo, Jorge Cardoso, and Michael Gerndt. "A Systematic Mapping Study in AIOps." International Conference on Service-Oriented Computing. Springer, Cham, 2020. ΑΓҾ༻

Slide 7

Slide 7 text

[Notaro ’20]: Notaro, P, Jorge C, and Michael G. "A Systematic Mapping Study in AIOps.” ICSOC. Springer, Cham, 2020. [Notaro ’20]: Fig.2ΑΓҾ༻ ITαʔϏεͷఏڙʹ͓͍ͯ๬·͘͠ ͳ͍ಈ࡞ʹରॲ͢Δํ๏ͷݚڀ ITαʔϏεΛ࠷దʹఏڙ͢ΔͨΊʹ ΤωϧΪʔɺܭࢉɺετϨʔδɺ࣌ ؒͷϦιʔεΛׂΓ౰ͯΔݚڀ AIOpsΛద༻ՄೳͳλεΫ

Slide 8

Slide 8 text

Ironies of Automation … the increased interest in human factors among engineers reflects the irony that the more advanced a control system is, so the more crucial may be the contribution of the human operator. ੍ޚγεςϜ͕ߴ౓ʹͳΕ͹ͳΔ΄ͲɺਓؒͷΦϖϨʔλͷߩݙ͕ΑΓॏཁʹͳΔ ͱ͍͏ൽ೑

Slide 9

Slide 9 text

Human-Centric AI ਓؒத৺ͷAI

Slide 10

Slide 10 text

Interactive AIOps AIʹΑΔ׬શࣗ཯ՔಇͷγεςϜΛ໨ࢦ͢ͷͰ͸ͳ͘   ਓؒΛத৺ͱͨ͠AIͱͷର࿩తΦϖϨʔγϣϯ

Slide 11

Slide 11 text

inter-active ITγεςϜͱਓؒͷ૒ํ͕ʮओମʯ

Slide 12

Slide 12 text

Interactive AIOpsͷཁ݅1: Experimentability ΦϖϨʔλʔ ITγεςϜ ڭ͑ΒΕΔ ΦϖϨʔλʔ͕ਖ਼ৗɾҟৗύ λʔϯΛITγεςϜʹ༩͑Δ ΦϖϨʔλʔ͕ҙਤతʹҟৗ Λൃੜͤ͞Δ ITγεςϜ͸ҟৗύλʔϯΛ ֶश͠ɺҟৗݕ஌ɾݪҼ෼ ੳɾࣗಈճ෮͕Մೳ

Slide 13

Slide 13 text

Interactive AIOpsͷཁ݅2: Explainability ΦϖϨʔλʔ ITγεςϜ ڭ͑ͨ಺༰Λ   આ໌Մೳ ITγεςϜ͸ֶशͨ͠ύλʔ ϯΛΦϖϨʔλʔʹग़ྗՄೳ ΦϖϨʔλʔ͸આ໌಺༰ʹෆ උ͕͋Ε͹ɺITγεςϜʹڭ ͑Δ͜ͱ͕Մೳ

Slide 14

Slide 14 text

Interactive AIOpsͷཁ݅3: ΦϖϨʔλʔ ITγεςϜ ະ஌΁ͷ൓Ԡ͕   Մೳ ڭ͍͑ͯͳ͍ɾܦݧ͍ͯ͠ͳ ͍͜ͱ͕ى͖͔ͨͲ͏͔ITγ εςϜ͕ݕ஌Մೳ ݕ஌ޙ͸ΦϖϨʔλʔʹ൑அ ΛҕͶΔ

Slide 15

Slide 15 text

Interactive AIOpsͷཁ݅4: ΦϖϨʔλʔ ITγεςϜ ଞͷγεςϜ͔Βֶ΂Δ ଞͷγεςϜ͔Βࣗ෼͕·ͩ ֶΜͰ͍ͳ͍͜ͱΛൃݟɾֶ शՄೳ ΦϖϨʔλʔ ITγεςϜ

Slide 16

Slide 16 text

• Ϋϥ΢υγεςϜͷৼΔ෣͍ΛϞσϧԽ • ϝτϦοΫɺϩάɺτϨʔεΛ୯ҰͷϞσϧͰѻ͏ • ۭؒߏ଄͕࣌ܥྻͰมԽ͍ͯ͘͠Πϝʔδ • GNNʢGraph Neural NetworkʣΛಈతʹมԽ͢Δάϥϑߏ଄ʹ ରԠͤ͞Δ ؔ࿈AIٕज़: ۭ࣌ؒϞσϧ

Slide 17

Slide 17 text

• ௨ৗ͸ɺ࣌ܥྻΛχϡʔϥϧωοτͰ֊૚Խ͢Δ • ࿈ଓͰ͋Δ࣌ܥྻ͕཭ࢄԽ͞ΕΔ • Neural ODE (2018 ) • NNͷ૚Λ࿈ଓԽͯ͠ѻ͏Ϟσϧ • ϝϞϦޮ཰ੑ޲্ɾύϥϝʔλ਺௿Լ • ਫ਼౓ͱ࣮ߦ଎౓ͷτϨʔυΦϑΛௐ੔Մೳ ؔ࿈AIٕज़: ࿈ଓ࣌ܥྻσʔλϞσϧ

Slide 18

Slide 18 text

• ݱ࣮ੈքͷ৘ใ͸ɺ௨ৗɺҟͳΔϞμϦςΟͱͯ͠ఏڙ͞Εɺ ͦΕͧΕҟͳΔ౷ܭతಛੑΛ΋ͭ • ҟͳΔϞμϦςΟؒͷؔ܎Λൃݟ͢Δ͜ͱ͸ඇৗʹॏཁ • ϚϧνϞʔμϧֶश͸ҟͳΔϞμϦςΟؒͷؔ܎Λൃݟ͢Δ͜ ͱ͸ඇৗʹॏཁ • ϝτϦοΫɾϩάɾτϨʔεΛҰݩతͳϞσϧͰදݱ ؔ࿈AIٕज़: ϚϧνϞʔμϧֶश

Slide 19

Slide 19 text

• eXplainable AIʢXAIʣ • ϩʔΧϧͳղऍʢSHAPͳͲʣ • Ϟσϧ͕༧ଌॲཧΛ࣮ߦͨ͠ͱ͖ʹɺ༧ଌॲཧͷࠜڌΛఏࣔ • άϩʔόϧͳղऍ • Ϟσϧࣗମ͕ԿΛॏཁࢹ͍ͯ͠Δ͔Λఏࣔ ؔ࿈AIٕज़: ਓ͕ؒղऍՄೳͳAI

Slide 20

Slide 20 text

• աڈʹֶशͨ͠λεΫͷ৘ใΛ৽ͨͳλεΫͷֶशʹ࠶ར༻· ͨ͸సૹ͢Δ • ܦݧʢҟৗͷൃੜʣ͕গͳ͍γεςϜʹɺܦݧ๛෋ͳγεςϜͷ ֶश݁ՌΛ࠶ར༻Մೳ ؔ࿈AIٕज़: సҠֶश

Slide 21

Slide 21 text

• ݱ୅ͷτϨϯυ͸ɺΦϖϨʔλʔ͕એݴͨ͠಺༰ʹITγεςϜΛ ௥ैͤ͞ΔํࣜͰ͋Δ • Interactive AIOps͸ΦϖϨʔγϣϯͷݪ఺ʹճؼ͍ͯ͠Δ UNIXͷର࿩؀ڥͱͷྨࣅੑ

Slide 22

Slide 22 text

௒ݸମܕσʔληϯλʔͱͷؔ࿈ • ௒ݸମܕDC͕ීٴ͢ΔͱɺDC਺͕രൃతʹ૿େ͢Δ • ӡ༻͕ࢸΔͱ͜Ζʹൃੜ͢Δ DevOpsͷۃக • ೝ஌ෛՙͷখ͍͞ର࿩తͳΦϖϨʔγϣϯ͕ཁٻ͞ΕΔ…͸ͣ

Slide 23

Slide 23 text

• 5೥લͷݚڀϏδϣϯͷ঺հ • ݱࡏࢥҊதͷݚڀϏδϣϯΛఏҊ • ίϯηϓτɿʮInteractive AIOpsʯ • ࠓޙͷ՝୊ • ϢʔβʔͷߦಈΛؚΊͨίϯηϓτͷఏࣔ ·ͱΊ