Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Sign up for free
Menu
Search
Features
All features
Private URLs
Password Protection
Custom URLS
Scheduled publishing
Remove Branding
Restrict embedding
Deck Collections
Notes
Features
All features
Private URLs
Password Protection
Custom URLS
Scheduled publishing
Remove Branding
Restrict embedding
Deck Collections
Notes
Explore
Featured decks
Featured speakers
Programming
Technology
Storyboards
Explore
Featured decks
Featured speakers
Programming
Technology
Storyboards
Pricing
Search
Sign in
Sign up for free
今日から始める Flood.io / fuka-taisaku-night-01
Search
kikunantoka
February 06, 2019
Technology
370
0
Share
Embed
Copy iframe code
Copy JS code
Copy link
Start on current slide
今日から始める Flood.io / fuka-taisaku-night-01
- ギフティでの負荷対策の重要性
- Flood.io について
- Flood.io の活用
- 30万RPM を耐えるためにやったこと
kikunantoka
February 06, 2019
More Decks by kikunantoka
See All by kikunantoka
個人開発しているサービスのインフラをAWSからGCPに載せ替えた話 💪 / kojin_kaihatsu_night_3
kikunantoka
0
1.3k
Gatsby.jsとCloud Functionsで毎週自動でコンテンツが更新され続けるフレームワーク比較サイトを作った話 / gotanda_js_13
kikunantoka
1
2.4k
Gatsby.jsとCloud Functionsで毎週自動でコンテンツが更新され続けるフレームワーク比較サイトを作った話 / frontend_night_1
kikunantoka
3
1.8k
サービスがゼロからN億円規模になるまに実践した7つのやっていき / 7_yatteiki_battle_conference_u30_2019
kikunantoka
1
1.4k
Gatsby.jsとNetlifyとの付き合い方 / gatsby-js-and-netlify
kikunantoka
3
650
Gatsby.jsで導入事例をバシバシ読めるSPAなLPを作った話 / gatsby-js-for-biz-lp
kikunantoka
2
750
Gatsby.jsで導入事例をバシバシ読めるSPAなLPを作った話 / gatsby-js-for-biz-lp
kikunantoka
1
1.9k
20万RPMを捌くRailsアプリケーションの作り方
kikunantoka
0
860
MVPに絞ったら個人開発でもちゃんとリリースできた話
kikunantoka
1
570
Other Decks in Technology
See All in Technology
2026_devsumi_ozono.pdf
o3
2
240
多層防御と最⼩権限で実現する、安全なAIエージェント設計パターン
lycorptech_jp
PRO
1
250
GoCon2026 - Open Source, Open World
sanposhiho
3
3.1k
アプリをもっと"iOSアプリっぽく"する小さな工夫 / Small Touches That Make Your App Feel More Like an iOS App
matsuji
1
330
Driving AI Adoption Using In-House GPUs to Serve Qwen
po3rin
2
480
AI時代のAPI品質を支えるガードレール / API Guardrails for API quality in the AI era
yokawasa
1
210
HRC_Frontend_Conference_Fukuoka_2026.pdf
ts020
0
360
GoにおけるFFIのこれまでとこれから
goccy
5
2.3k
When Does a Local Qwen Start to Break
morshoto
0
200
DEFCON34-Write-up_HYCu-MYCu
daikiokazaki
0
140
アプリログインとWeb認証基盤をつなぐ ASWebAuthenticationSession 作法
shimastripe
1
190
WAF 運用改善の承認サイクル/SRE_BizReach_MIXI_1
visional_engineering_and_design
2
420
Featured
See All Featured
A Tale of Four Properties
chriscoyier
163
24k
How to build an LLM SEO readiness audit: a practical framework
nmsamuel
1
890
Claude Code どこまでも/ Claude Code Everywhere
nwiizo
67
58k
Building Adaptive Systems
keathley
44
3.2k
Designing for Performance
lara
611
70k
Marketing Yourself as an Engineer | Alaka | Gurzu
gurzu
0
300
Statistics for Hackers
jakevdp
799
230k
Redefining SEO in the New Era of Traffic Generation
szymonslowik
1
410
The Language of Interfaces
destraynor
162
27k
Dealing with People You Can't Stand - Big Design 2015
cassininazir
367
27k
Building AI with AI
inesmontani
PRO
1
1.2k
A better future with KSS
kneath
240
18k
Transcript
ࠓ͔Β࢝ΊΔ Flood.io גࣜձࣾΪϑςΟ ٠ ࢙و ෛՙରࡦ Night #1 @ΪϑςΟ on
2018/02/06
ࣗݾհ about: name: Fumitaka Kikukawa twitter: @kikunantoka job: engineer work_at:
giftee Inc. // giftee is a good company. url: https://kikunantoka.com
ࣗݾհ
ࣗݾհ https://nakamy.com
͍͑ͨ͜ͱ • ΪϑςΟͰͷෛՙରࡦͷॏཁੑ • Flood.io ʹ͍ͭͯ • Flood.io ͷ׆༻ •
30ສRPM Λ͑ΔͨΊʹͬͨ͜ͱ
ΪϑςΟͰͷෛՙରࡦ ͷॏཁੑ
ࣄۀ༰
ࣄۀ༰
ࣄۀ༰
giftee for Buisness
୲͍ͯ͠ΔϓϩμΫτ ʢࣾͰ($1ͱུ͞Εͩ͢ʜʣ HJGUFFΠϯελϯτΟϯGPS5XJUUFSͷڧΈᶃ நબ͔Βܠͷఏڙ·ͰҰؾ௨؏Ͱ͝ఏڙ நબγεςϜ ʢΠϯελϯτΟϯʣ σδλϧΪϑτͷఏڙ ʢछྨҎ্ͷܠʣ ϦΞϧλΠϜʹநબɺ ͦͷͰܠΛ༩
LINEೝূͷύλʔϯఏڙ HJGUFFΠϯελϯτΟϯGPS-*/&ͷڧΈᶃ ©2018 gi)ee Inc. all rights reserved J நબ͔Βܠͷఏڙ·ͰҰؾ௨؏Ͱ͝ఏڙ
நબγεςϜ ʢΠϯελϯτΟϯʣ σδλϧΪϑτͷఏڙ ʢछྨҎ্ͷܠʣ ϦΞϧλΠϜʹநબɺ ͦͷͰܠΛ༩
ΊͬͪΌΞΫηεདྷΔ • ༑ͩͪ 2300ສͷΞΧϯτͰΩϟϯϖʔϯΛ࣮ࢪͨ݁͠Ռ ʊਓਓਓਓਓਓਓਓਓਓਓʊ ʼɹ࠷େ 30ສRPMɹʻ ʉY^Y^Y^Y^Y^Y^Y^Y^ʉ
Flood.io ʹ͍ͭͯ
Flood.io ͱ • ΫϥυϕʔεͷϩʔυςεταʔϏε
Flood.io ͱ • Selenium, JMeter ( + ruby-jmeter) , Gatling
• ͖ͳςετπʔϧΛ͏͜ͱ͕Ͱ͖Δ • طʹGatlingͷςετίʔυ͕͋ͬͨͷͰɺGatlingΛ࠾༻
Flood.io ͱ • Ձ֨ • 50 Node Hours Ͱ 33,000ԁఔʢैྔׂҾ͋Γʣ
Flood.io ͷྑ͍ • Ϋϥυ্Ͱؾܰʹϩʔυςετ͕Ͱ͖Δ • ϊʔυͷεέʔϧΞοϓ͕؆୯ • ݁Ռ͕ϏδϡΞϥΠζ͞ΕΔ • Web্Ͱ݁Ռͱςετέʔεͷཧ͕Ͱ͖Δ
• ςετ݁ՌγΣΞ༻ͷϦϯΫͰڞ༗Ͱ͖Δ
Flood.io ͷѱ͍ • ϩάͷอ༗ظ͕ؒ1ϲ݄ఔ • ςετέʔεͱΤϥʔݪҼͷηοτͰཧ͍ͨ͠߹ʹෆ ศ
σϞ
Flood.io ͷ׆༻
New Relic ͰϘτϧωοΫΛಛఆ • Flood.io ͰෛՙΛ͔͚ͨ࣌ͷ༷ࢠΛϞχλϦϯά͢Δ
ෛՙݕূΛ܁Γฦ͢ • ϩʔυςετΛ͢Δ • ϘτϧωοΫ͕ݟ͔ͭΔ • ϘτϧωοΫΛվળ͢Δ • Λ܁Γฦ͢
ෛՙݕূΛ܁Γฦ͢ • ϩοΫॲཧʹΑͬͯɺλΠϜΞτ • ϩοΫॲཧͷൣғͷվળ • ίωΫγϣϯϓʔϧ͕Γͳ͘ͳΔ • DBͷઃఆͷमਖ਼ •
RDSʹଓͰ͖ͳ͘ͳΔ • ΠϯελϯελΠϓͷมߋ
ϘτϧωοΫʹͳΓ͍͢ϙΠϯτ • σʔλ͕૿͑Δ͜ͱʹΑͬͯੑೳྼԽ͠ͳ͍͔ • εϩʔΫΤϦΛ͍͛ͯͳ͍͔ • ϩοΫॲཧͷൣғ͕͗͢ͳ͍͔ • σουϩοΫ͍ͯ͠ͳ͍͔ •
ແବͳΠϯελϯεΛੜ͍ͯ͠ͳ͍͔ • ੩తϑΝΠϧCDN৴͢Δ
ҙ • AWSࣄલʹϩʔυςετ͢ΔࢫΛ͓͑ͯ͘ • ಉҰIP͔ΒҰఆҎ্ΞΫηε͕͋Δͱ߈ܸͱΈͳ͞ΕΔ • AWS LoftͷαϙʔτΤϯδχΞਃग़ͨ͠΄͏͕ྑ͍ ͱͷݟղ
30ສRPM Λ͑ΔͨΊʹ ͬͨ͜ͱ
εέʔϧΞοϓͰ͖ΔΑ͏ʹ͓ͯ͘͠ • Elastic Beanstalk
εέʔϧΞοϓͰ͖ΔΑ͏ʹ͓ͯ͘͠ • Amazon Aurora • RDS for MySQL͔ΒҠߦͨ͠ • ϑΣΠϧΦʔόʔػೳΛ͏͜ͱͰɺΠϯελϯελΠϓ
ͷมߋ࣌ͷμϯλΠϜ͕5ඵఔʹ • ॻ͖ࠐΈIOPS͔Βͷ։์
DBͷઃఆΛݟ͢ • DB ͷ Pool Puma ͷ Thread ɺWorker
ద͔ • DB ͷ Pool -> Puma ͷεϨου • Puma ͷ Worker -> CPUͷίΞ • Puma ͷ Thread -> CPU༻ͱ૬ஊ
࠷৽ͷΠϯελϯελΠϓΛ͏ ໊લ W$16 3". $16ΫϨδοτ࣌ؒ ྉۚ࣌ؒ UTNBMM
64% UTNBMM 64% • ίεύ͕ྑ͍ • Puma ͷ Worker Λ 2 ʹͰ͖Δ -> 2ഒͷεϨου • ͨͩ͠ɺAWSଆͰϦιʔε͕Γͳ͘ͳΔϦεΫ͕͋Δ
WebαʔόΛཱͯ·͘Ε͍͍͍ͬͯ͏Ͱͳ͍ • db.r4.16xlargeͷ߹ • 32 Threads x 2 Workers •
1͋ͨΓ 64 Threads • 6000 / 64 = 93.75 • 92͙Β͍͕ݶքʂ
ͪΌΜͱਓΛೖΕΔ • 1ਓͩͱӡ༻ਏ͍ • 2ਓ͍Εɺਏ͍͜ͱʹɺتͼ2ഒʹ • ͓ۚΛՔ͙ • Λ্͛Δ
ίϚʔγϟϧ • We are hiring!!! - Ұॹʹಇؒ͘Λืूதʂ ձһສਓಥഁʂ ࠃ/PͷΧδϡΞϧΪϑταʔϏε