Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
Web サービスの信頼性と運用の自動化について / iot40-rrreeeyyy
Search
rrreeeyyy
March 08, 2018
Technology
12
7.6k
Web サービスの信頼性と運用の自動化について / iot40-rrreeeyyy
第40回インターネットと運用技術研究発表会の招待講演で『Web サービスの信頼性と運用の自動化について』というタイトルで発表しました。
rrreeeyyy
March 08, 2018
Tweet
Share
More Decks by rrreeeyyy
See All by rrreeeyyy
Incident Response Practices: Waroom's Features and Future Challenges
rrreeeyyy
0
180
An Efficient Incident Response Training with AI / SRE NEXT 2024 Sponsor Session
rrreeeyyy
1
3.6k
カンファレンスから見る SRE トレンド 2024 / SRE Trends from Conferences in 2024 #SRE_Findy
rrreeeyyy
4
2.2k
信頼性の育て方 / mackerel-meetup-15
rrreeeyyy
10
2.4k
SRE の歩き方・進め方 / sre-walk-through-procedure
rrreeeyyy
0
8.6k
「信頼性」を保ちつつ大規模サービスをリニューアルする / cookpad-tech-kitchen-service-embedded-sres
rrreeeyyy
11
12k
Cookpad and Prometheus
rrreeeyyy
6
20k
SRE-Lounge-8-Cookpad-Microservice-Architecture-Overview
rrreeeyyy
5
5.3k
A survey of anomaly detection methodologies for web system
rrreeeyyy
5
1.2k
Other Decks in Technology
See All in Technology
OpenAIの蒸留機能(Model Distillation)を使用して運用中のLLMのコストを削減する取り組み
pharma_x_tech
4
560
KubeCon NA 2024 Recap: How to Move from Ingress to Gateway API with Minimal Hassle
ysakotch
0
200
どちらを使う?GitHub or Azure DevOps Ver. 24H2
kkamegawa
0
820
MLOps の現場から
asei
6
650
サービスでLLMを採用したばっかりに振り回され続けたこの一年のあれやこれや
segavvy
2
470
ブラックフライデーで購入したPixel9で、Gemini Nanoを動かしてみた
marchin1989
1
540
GitHub Copilot のテクニック集/GitHub Copilot Techniques
rayuron
37
14k
DevFest 2024 Incheon / Songdo - Compose UI 조합 심화
wisemuji
0
110
NW-JAWS #14 re:Invent 2024(予選落ち含)で 発表された推しアップデートについて
nagisa53
0
270
社外コミュニティで学び社内に活かす共に学ぶプロジェクトの実践/backlogworld2024
nishiuma
0
260
podman_update_2024-12
orimanabu
1
280
Microsoft Azure全冠になってみた ~アレを使い倒した者が試験を制す!?~/Obtained all Microsoft Azure certifications Those who use "that" to the full will win the exam! ?
yuj1osm
2
110
Featured
See All Featured
Scaling GitHub
holman
458
140k
Making the Leap to Tech Lead
cromwellryan
133
9k
Automating Front-end Workflow
addyosmani
1366
200k
I Don’t Have Time: Getting Over the Fear to Launch Your Podcast
jcasabona
29
2k
GitHub's CSS Performance
jonrohan
1030
460k
Optimizing for Happiness
mojombo
376
70k
Why You Should Never Use an ORM
jnunemaker
PRO
54
9.1k
RailsConf & Balkan Ruby 2019: The Past, Present, and Future of Rails at GitHub
eileencodes
132
33k
How to Create Impact in a Changing Tech Landscape [PerfNow 2023]
tammyeverts
48
2.2k
Git: the NoSQL Database
bkeepers
PRO
427
64k
The Cost Of JavaScript in 2023
addyosmani
45
7k
How STYLIGHT went responsive
nonsquared
95
5.2k
Transcript
Web αʔϏεͷ৴པੑͱ ӡ༻ͷࣗಈԽʹ͍ͭͯ ΫοΫύουגࣜձࣾ ΠϯϑϥετϥΫνϟʔ෦ SRE άϧʔϓ ٢ ཽଠ (
@rrreeeyyy ) ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy ) 1
Agenda • Site Reliability Engineering (SRE) ʹ͍ͭͯ • SRE ͱ
• ӡ༻ͷݱͱӡ༻ٕज़ݚڀ • ӡ༻ͷࣗಈԽʹ͍ͭͯ • ࣌ܥྻσʔλͷॲཧͱҟৗݕʹ͍ͭͯ ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy ) 2
Me !SSSFFFZZZ !SSSFFFZZZ IUUQTSSSFFFZZZDPN :PTIJLBXB3ZPUB ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) |
Yoshikawa Ryota ( @rrreeeyyy ) 3
Me • Yoshikawa Ryota ( @rrreeeyyy [reɪ] ) • ిؾ௨৴େֶ
(2010/04 ʙ 2014/03) • ใཧֶ෦ ใɾ௨৴ֶՊ • גࣜձࣾϋʔτϏʔπ (2010/11ʙ2016/12) • MSP ࣄۀ෦ ӡ༻ελοϑɾӡ༻ΤϯδχΞ • ΫοΫύουגࣜձࣾ (2017/01 ʙ) • ΠϯϑϥετϥΫνϟʔ෦ SRE άϧʔϓ ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy ) 4
Me • ిؾ௨৴େֶ (2010/04 ʙ 2014/03) • mruby Λ༻͍ͨLinux ϩʔυόϥϯαΠϯλϑΣʔεͷ࣮Ͱଔۀ
• גࣜձࣾϋʔτϏʔπ (2010/11ʙ2016/12) • ʙઍنͷӡ༻ɾߏஙɾཧΛܦݧ • ϝσΟΞɾήʔϜɾEC ͷ༷ʑͳͷӡ༻Λܦݧ • ΫοΫύουגࣜձࣾ (2017/01 ʙ) • ΠϯϑϥετϥΫνϟʔ෦ SRE άϧʔϓ • ઍنͷγεςϜͷ৴པੑʹؔΘΔ෦શͯΛ୲ ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy ) 5
ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy )
6
Site Reliability Engineering ͱ • αΠτͷ৴པੑͷશͯʹൣғΛ࣋ͭɺͱ͢ΔΤϯδχΞϦϯάख๏ • ಛʹɺιϑτΣΞɾγεςϜΤϯδχΞϦϯάΛ༻͍ͯ՝Λղܾ͢Δ • ྫ͑ɺύϑΥʔϚϯε
• ύϑΥʔϚϯεʹ͕͋Δίʔυ SRE ͕ࣗίʔυΛॻ͍ͯमਖ਼͢Δ • ྫ͑ɺεέʔϥϏϦςΟɾΞϕΠϥϏϦςΟ • ਓؒʹΑΔखಈͷ࡞ۀΛίʔυʹΑΓࣗಈԽ͢Δ • نΞΫηε͕૿େͯ҆͠ఆͯ͠αʔϏεΛఏڙͰ͖Δঢ়ଶʹ͢Δ • ΠϯϑϥΤϯδχΞͷൣғ͕૿େͨ͠தͰ͋Δछɺ৬Λ࠶ఆٛͨ͠ͱଊ͍͑ͯΔ ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy ) 7
SRE ͷ׆ಈ SRE ͷ׆ಈେ͖͚ͯ࣍͘ͷ 4 ͭʹྨͰ͖Δͱ͞Ε͍ͯΔ • So#ware Engineering •
Systems Engineering • Toil • Overhead ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy ) 8
SRE ͷ׆ಈ • So#ware Engineering • ίʔυͷ࡞मਖ਼, ͦΕʹؔ࿈͢ΔઃܭυΩϡϝϯςʔγϣϯͷ࡞ۀͳͲ • ࣗಈԽεΫϦϓτͷ࡞,
πʔϧϑϨʔϜϫʔΫͷ࡞ ... • αʔϏεͷ৴པੑΛߴΊΔػೳͷՃ ... • Systems Engineering • 1 ճͷ࡞ۀͰӬଓతͳվળΛੜΉ࣍ͷΑ͏ͳ࡞ۀ • ຊ൪γεςϜͷઃఆมߋ ... • ։ൃνʔϜͷαʔϏεΞʔΩςΫνϟઃܭ ... ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy ) 9
SRE ͷ׆ಈ • Toil • αʔϏεΛՔಇͤ͞ΔͨΊ͚ͩʹ܁Γฦ͠ߦΘΕΔखಈͰͷ࡞ۀ • ظతͳՁΛ࣋ͨͳ͍, αʔϏεͷʹରͯ͠ O(n)
Ͱ͋Δ ... • ͭ·Βͳ͍ࣄͰظతͳՁ͕͋Ε Toil Ͱͳ͍ • Overhead • αʔϏεΛՔಇͤ͞Δ͜ͱʹ݁͠ͳ͍ཧతͳ࡞ۀ • ࠾༻, ϛʔςΟϯά, ධՁ, τϨʔχϯά ... ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy ) 10
ӡ༻ͷݱͱӡ༻ٕज़ݚڀ • ΤϯδχΞϦϯάΛͪΌΜͱΔࣄ͕ݱϨϕϧͰධՁ͞ΕΔΑ͏ʹ • ͜ͷจ຺ͷҰͭͱͯ͠ SRE ͷΑ͏ͳख๏ɾ৬छ͕ग़͖ͯͨͱ͍͏ཧղ • จΛॻ͍ͨΓಡΜͰཧղɾ࣮͢Δࣄ͕ධՁ͞ΕΔΑ͏ʹͳ͖ͬͯͨ •
࣮ࡍͷӡ༻ͷݱͰಡΜͰཧղ͢Δࣄඞཁʹͳ͖ͬͯͨ • ٕज़બఆͷࢦඪͷҰͭͱͯ͠ϛυϧΣΞͷڍಈͷཧղͳͲ • ͍ΘΏΔݚڀ։ൃͷΑ͏ͳ෦Ͱͳͯ͘ৗతʹߦ͏Α͏ʹ • ଟ͘ͷٕज़اۀͰӡ༻ɾϛυϧΣΞͷจ͕ଟ͘ग़ͯ͘ΔΑ͏ʹ ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy ) 11
େखΫϥυϕϯμͷจ • Google, Amazon, Facebook, Ne3lix... • ଞʹେখؔΘΒ༷ͣʑͳاۀ͔Βจ͕ग़͍ͯΔ • USENIX
ͷΧϯϑΝϨϯεͰٕज़اۀ͕ੵۃతʹൃද • FAST, OSDI, Security ... • ΫϥυαʔϏεͷ෦࣮ϛυϧΣΞؔΘΒͣ • ΫϥυαʔϏεͷڍಈΛਖ਼͘͠ཧղ͢ΔͨΊʹจΛಡΉ • ৽͍͠ΞʔΩςΫνϟɾϛυϧΣΞͷબఆͷͨΊʹจΛಡΉ ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy ) 12
ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy )
13
ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy )
14
ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy )
15
ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy )
16
ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy )
17
จͷ༰Λ࣮ӡ༻Ͱ͑ΔϛυϧΣΞ • HashiCorp • Vagrant, Terraform Λ࡞͍ͬͯΔϕϯνϟʔ • จͷ༰Λ࣮ӡ༻Ͱ͍͍͢ܗʹམͱ͠ࠐΜͰ͍Δ •
Serf, Consul, Vault • ͜ΕΒͷϛυϧΣΞΛਖ਼͘͠ཧղͯ͠ӡ༻͢ΔͨΊʹจΛಡΉ • ͳͥ Consul ͷϊʔυ࠷ 3 ඞཁͳͷ͔ʁ ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy ) 18
ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy )
19
ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy )
20
ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy )
21
ΫοΫύουʹ͓͚Δ SRE ͷ׆ಈ ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota
( @rrreeeyyy ) 22
ΫοΫύουʹ͓͚Δ SRE ͷ׆ಈ • SRE ͱͯ͠ͷ׆ಈΛΫοΫύουͰ༷ʑߦ͍ͬͯΔ • ΫοΫύουͷ SRE ͷ׆ಈΛ͍͔ͭ͘հ
• ৽نαʔϏεʹର͢Δ SRE ͷಈ͖ • ϞχλϦϯάͷվળ • ϩʔυόϥϯγϯάͷվળ • ৽ଔݚम • ... ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy ) 23
ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy )
24
ΫοΫύουʹ͓͚Δ SRE ͷ׆ಈ • SRE ͱͯ͠ͷ׆ಈΛΫοΫύουͰ༷ʑߦ͍ͬͯΔ • ΫοΫύουͷ SRE ͷ׆ಈΛ͍͔ͭ͘հ
• ৽نαʔϏεʹର͢Δ SRE ͷಈ͖ • ϞχλϦϯάͷվળ • ϩʔυόϥϯγϯάͷվળ • ৽ଔݚम • ... ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy ) 25
ϞχλϦϯάͷվળ • େྔͷσʔλΛೖɾॲཧ͢ΔΑ͏ͳ࣌ܥྻσʔλϕʔεͷج൫͖ͬͯͨ • ΫοΫύουͰ Prometheus Λ TSDB ͱͯ͠࠾༻ •
Long-term storage ͱͯ͠ S3 ΛόοΫΤϯυͱͯ͠ݕ౼த • औಘͰ͖ΔϝτϦΫεͷղ૾ͷ্ɾछྨͷ૿ڧ • 15 ඵ୯ҐͷϝτϦΫεऔಘ͕ग़དྷΔΑ͏ʹͳ͍ͬͯΔ • ϚΠΫϩαʔϏεͷαʔϏεؒͷ࿈ܞͷ༷ࢠՄࢹԽΛݕ౼த ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy ) 26
ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy )
27
ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy )
28
ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy )
29
ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy )
30
࣌ܥྻσʔλͷॲཧʹ͍ͭͯ • ͜Ε·Ͱ΄ͱΜͲ੩తͳᮢͰԿͱ͔͍ͯͨ͠ঢ়ଶ • ͜Ε͔Β࣌ܥྻσʔλʹ؆୯ͳ౷ܭతॲཧ͕ࢪͤΔΑ͏ʹ • ઢܗճؼ, ࢦฏԽ, ҠಈฏۉͳͲ... •
ࣄલʹى͖͏Δҟৗ͕͔͍ͬͯΔ࣌ʹΑ͘ಈ͘ • Πϯγσϯτোͷ "ੵΈॏͶ" ͰݕͰ͖ΔΑ͏ʹ • ࣮ࡍʹऔΕΔϝτϦΫε͕૿͑ҟৗ࣌ʹͲ͏ͳΔ͔༧ଌࠔͳࣄ ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy ) 31
ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy )
32
ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy )
33
࣌ܥྻσʔλͱҟৗݕ • ࣌ܥྻσʔλͷҟৗ(֎Ε)ݕ༷ʑͳख๏͕طʹݚڀ͞Ε͖͍ͯͯΔ • Gaussian Mixture Model, Isola5on forest ͳͲͳͲ...
• ࣮ࡍͷӡ༻Ͱग़ݱ͢ΔϝτϦΫεͰඇઢܗͳཁૉ͕ଟ͘ग़ݱ͢Δ • ैདྷͷख๏ෳࡶͳඇઢܗมͳͲͷදݱʹऑ͍͜ͱ͕͋ͬͨ • ࠷ۙͷख๏Ͱ CNN Λ༻͍ͯҟৗݕߦ͏Α͏ͳख๏͕ొͨ͠ • ෳͷೖྗ͕͋ΔΑ͏ͳ૬ؔҟৗͰൺֱత؆୯ʹ༧ଌ͢Δࣄ͕Մೳ ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy ) 34
ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy )
35
ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy )
36
ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy )
37
LSTM ʹΑΔ࣌ܥྻσʔλͷҟৗݕ • LSTM Λ༻͍ͯਖ਼ৗͳ࣌ܥྻσʔλͷֶशΛߦ͏ • ߦֶͬͨशʹج͍ͯɺ࣍ͷσʔλͷ༧ଌΛߦ͏ • ࣮ࡍͷσʔλͱ༧ଌΛൺֱ͢Δ͜ͱͰҟৗΛଌఆ͢Δ •
ֶश͢Δσʔλͷೖྗͱͯ͠ϝτϦΫεҎ֎ʹαʔόͷίϯςΩετΛೖΕΔ • ྫ͑σϓϩΠதɾόον࣮ߦதɾΦʔτεέʔϧத... • ͦͷ࣌ͷαʔόͷίϯςΩετΛؚΊͯ༧ଌΛ࡞ΕΔ • ΑΓαʔόͷঢ়گΛՃຯͨ͠ҟৗΛଌΔ͜ͱ͕ग़དྷΔ ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy ) 38
༧ଌରʹͳΔ࣌ܥྻσʔλͷྫ timestamp,cpu,deploying,autoscaling,batching 2018-03-05 12:00:00,80,1,0,0 2018-03-05 12:00:15,81,1,0,0 2018-03-05 12:00:30,82,1,0,0 2018-03-05 12:00:45,82,1,1,0
2018-03-05 12:01:00,75,1,1,0 2018-03-05 12:01:15,60,1,1,0 2018-03-05 12:01:30,55,1,1,0 2018-03-05 12:01:45,40,0,1,0 2018-03-05 12:02:00,35,0,0,0 • ࣮ࡍʹ Prometheus ͷϥϕϧػೳΛ༻্͍ͯهͷίϯςΩετΛೖྗ ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy ) 39
Keras Λ༻͍࣮ͨɾ࣮ݧ • ϊΠζΛؚ·ͤͨਖ਼ݭΛ LSTM Ͱֶशͤ͞Δ • ݩͱͳͬͨਖ਼ݭʹҟৗΛؚ·ͤΔ • ҟৗͷ͋ΔܗͱֶशࡁΈͷϞσϧ͔Βੜ͞ΕͨܗͷೋޡࠩΛऔΔ
• ߋʹ࣮ࡍʹલϖʔδͷ࣌ܥྻσʔλΛֶशͯ͠༧ଌΛग़ͯ͠ΈΔ • ࣮ࡍͷσʔλʹݶΓͳ͍ۙ͘ܗͷ toy-data ͱͯ͠ར༻ • CPU ༻ʹؔͯ͠ 0-1 ͷൣғʹਖ਼نԽֶͯ͠शΛऩଋͤ͞Δ • ز͔ͭͷ CPU ༻ɾσϓϩΠपظͷϞσϧΛֶशͤͯ͞༧ଌͤͯ͞ΈΔ ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy ) 40
ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy )
41
ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy )
42
ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy )
43
ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy )
44
·ͱΊ • SRE ͱ͍͏ΤϯδχΞϦϯάख๏ʹ͍ͭͯհ • ΠϯϑϥΤϯδχΞͷ৬͕͕͖ͬͯͨ • αΠτͷ৴པੑࣗମʹ৬Λ࣋ͭ͜ͱΛ໌֬ʹ͢Δ SRE ͕ొ
• αΠτͷ৴པੑΛकΔͨΊʹ "ਖ਼͘͠ΤϯδχΞϦϯάΛ͢Δ" ࣄ͕ॏཁ • ΫοΫύουͰͷ SRE ͷऔΓΈͱͯ͠ϞχλϦϯάͷվળΛհ • ࣌ܥྻσʔλΛରʹอଘɾղੳ͕ߦ͑Δج൫͕͖ͬͯͨ • ࣍ͳΔεςʔδͱͯ࣌͠ܥྻσʔλͷҟৗݕʹ͍༷ͭͯʑͳख๏Λࡧத ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy ) 45
ࢀߟࢿྉ • h#ps:/ /landing.google.com/sre/book.html • h#ps:/ /www.allthingsdistributed.com/files/p1041-verbitski.pdf • h#p:/ /lamport.azurewebsites.net/tla/formal-methods-amazon.pdf
• h#ps:/ /staAc.googleusercontent.com/media/research.google.com/en/ /archive/spanner-osdi2012.pdf • h#p:/ /techlife.cookpad.com/entry/Ameseries-database-001 • h#p:/ /www.vldb.org/pvldb/vol8/p1816-teller.pdf • h#ps:/ /qiita.com/rrreeeyyy/items/6bbd1f09c96e2af4a8ce • h#p:/ /ramcloud.stanford.edu/raH.pdf • h#p:/ /www.cs.cornell.edu/~asdas/research/dsn02-swim.pdf • h#ps:/ /speakerdeck.com/rrreeeyyy/jK-2017-site-reliability-engineering • h#ps:/ /speakerdeck.com/yuukit/the-study-of-Ame-series-database-for-server-monitoring • h#ps:/ /speakerdeck.com/rrreeeyyy/introducAon-to-prometheus-pracAce • h#ps:/ /github.com/NeKlix/vizceral • h#ps:/ /github.com/nghialv/promviz ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy ) 46
PR ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy
) 47
We are hiring! • ΫοΫύουͰຊհͨ͠Α͏ͳऔΓΈΛߦ͍ͬͯ·͢ • ·ͩ·ͩղܾ͍ͨ͠ɾऔΓΈ͍ͨ՝͕ͨ͘͞Μ͋Δ • ָ͠Έͳ͕ΒҰॹʹΕΔϝϯόʔΛืू͍ͯ͠·͢ →
h$ps:/ /cookpad.jobs ୈ 40 ճΠϯλʔωοτͱӡ༻ٕज़ݚڀൃදձ (2018/03/05) | Yoshikawa Ryota ( @rrreeeyyy ) 48