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
KARTE を支えるマルチプラットフォームインフラ監視 /karte-multi-platfo...
Search
Daiki Matsui
January 20, 2017
Technology
16k
2
Share
Embed
Copy iframe code
Copy JS code
Copy link
Start on current slide
KARTE を支えるマルチプラットフォームインフラ監視 /karte-multi-platform-monitoring
【freee × プレイド】Tech Meetup 〜インフラ監視編〜
Daiki Matsui
January 20, 2017
More Decks by Daiki Matsui
See All by Daiki Matsui
The Fourth Age of SRE
ikemonn
0
900
マルチクラウドな大規模サービスKARTEを支えるDatadog/multi-cloud monitoring with Datadog
ikemonn
0
2.6k
Multicloud deploy with Spinnaker / builderscon tokyo 2018
ikemonn
2
4.9k
Node Apps Performance Tuning
ikemonn
0
150
Other Decks in Technology
See All in Technology
AIをフル活用してオンコール機能のプロトタイプを2日で作った話 / Building an AI-Powered On-Call Prototype in Just Two Days
nari_ex
0
160
WebGIS AI Agentの紹介
_shimizu
0
600
プライバシー保護の理論と実践
lycorptech_jp
PRO
1
200
5分でわかる Amazon Connect_20260608
hwangbyeonghun
0
150
飲食店もAIで。レジ締めやハンディシステムをつくってる話 / Using AI for restaurant management
vtryo
0
210
サイバーエージェントにおけるAI推進戦略と変革への取り組み
shotatsuge
0
640
“全部コピーしない”ファイルデータの活用 : — FSx for ONTAP × S3 Tables × Icebergで作るメタデータカタログ
yoshiki0705
0
130
製造現場での生成AIの活用、およびエージェントAIの実装のあり方、AVEVAの取り組み
iotcomjpadmin
0
210
徹底討論!ECS vs EKS!
daitak
3
1.8k
「ビジネスがわかるエンジニア」とは何か?
ryooob
0
440
#エンジニアBooks 30分でわかる 「技術記事を書く技術」 / engineer-books 2026-06-30
jnchito
1
150
技術・能力を向上する原理原則 #きのこセッションa #きのこ2026
bash0c7
0
190
Featured
See All Featured
Primal Persuasion: How to Engage the Brain for Learning That Lasts
tmiket
0
380
Dominate Local Search Results - an insider guide to GBP, reviews, and Local SEO
greggifford
PRO
0
200
Mozcon NYC 2025: Stop Losing SEO Traffic
samtorres
1
260
The #1 spot is gone: here's how to win anyway
tamaranovitovic
3
1.1k
Designing for Performance
lara
611
70k
Paper Plane
katiecoart
PRO
1
52k
Building Applications with DynamoDB
mza
96
7.1k
<Decoding/> the Language of Devs - We Love SEO 2024
nikkihalliwell
1
260
Kristin Tynski - Automating Marketing Tasks With AI
techseoconnect
PRO
0
280
Responsive Adventures: Dirty Tricks From The Dark Corners of Front-End
smashingmag
254
22k
Speed Design
sergeychernyshev
33
1.9k
Avoiding the “Bad Training, Faster” Trap in the Age of AI
tmiket
0
190
Transcript
KARTEΛࢧ͑Δ ϚϧνϓϥοτϑΥʔϜΠϯϑϥࢹ @ikemonn
@ikemonn ΠϯϑϥΤϯδχΞ 201511݄ʹjoin
࣍ 1. αʔϏεͷհ 2. ࢹͷత 3. Πϯϑϥߏ 4. ࢹͷํ๏ 5.
Ξϥʔτ 6. ϋϚͬͨ͜ͱ 7. ·ͱΊ
࣍ 1. αʔϏεͷհ 2. ࢹͷత 3. Πϯϑϥߏ 4. ࢹͷํ๏ 5.
Ξϥʔτ 6. ϋϚͬͨ͜ͱ 7. ·ͱΊ
Σϒ٬ϓϥοτϑΥʔϜ KARTEΛ։ൃ͍ͯ͠·͢
Σϒ٬ϓϥοτϑΥʔϜ KARTEΛ։ൃ͍ͯ͠·͢
ΠϯϑϥΤϯδχΞ2ਓ
࣍ 1. αʔϏεͷհ 2. ࢹͷత 3. Πϯϑϥߏ 4. ࢹͷํ๏ 5.
Ξϥʔτ 6. ϋϚͬͨ͜ͱ 7. ·ͱΊ
ʮશһͰյ͠ͳ͕ΒਐΉʯ ͱ͍͏จԽΛࢧ͑Δ
Startup։ൃεϐʔυ͕ॏཁ
৽͍͠αʔϏεɺݴޠΛੵۃతʹ ಋೖͰ͖ΔΑ͏ʹ
ރΕٕͨज़Ͱͳ͍ͷͰɺ ใ͕͋·Γݟ͔ͭΒͳ͍
ଟগͷࣦഊڐ༰ & ࣦഊͨ࣌͠ʹ͙͢ؾ͚Δ
ʮશһͰյ͠ͳ͕ΒਐΉʯจԽΛ ࢧ͑ΔͨΊʹࢹͷڧԽ !
࣍ 1. αʔϏεͷհ 2. ࢹͷత 3. Πϯϑϥߏ 4. ࢹͷํ๏ 5.
Ξϥʔτ 6. ϋϚͬͨ͜ͱ 7. ·ͱΊ
ϚϧνϓϥοτϑΥʔϜ ৽͍͠αʔϏεͷੵۃར༻
None
࣍ 1. αʔϏεͷհ 2. ࢹͷత 3. Πϯϑϥߏ 4. ࢹͷํ๏ 5.
Ξϥʔτ 6. ϋϚͬͨ͜ͱ 7. ·ͱΊ
ࢹํ๏
ผϓϥοτϑΥʔϜͷΠϯελϯε ಉ࢜ͷϝτϦοΫൺֱ͍ͨ͠ ֤ϓϥοτϑΥʔϜʹґଘͨ͠ϝτ ϦοΫૹ৴ͷίʔυΛॻ͖ͨ͘ͳ͍ ▪Πϯϑϥଆͷ՝ ▪ΞϓϦଆͷ՝
CloudWatch ৭ʑͳཧը໘Λݟͨ͘ͳ͍ http://www.irasutoya.com/2016/06/blog-post_147.html https://support.draw.io/pages/viewpage.action?pageId=1671326 https://en.wikipedia.org/wiki/File:Google_Stackdriver_logo.svg https://www.runscope.com/
جຊతʹDatadogͰҰݩཧ
ԿΛࢹ͢Δͷ͔ʁ
σʔλͷૹ৴ݩ σʔλͷૹ৴ઌ ௨ઌ CloudWatch Datadog Agent library agent https://www.datadoghq.com/ https://nz.pinterest.com/pin/296393219208794919/
https://bugsnag.com/ https://logdna.com/ https://www.pagerduty.com/resources/logo/
Πϯϑϥͷࢹ
σʔλͷૹ৴ݩ σʔλͷૹ৴ઌ ௨ઌ CloudWatch Datadog Agent library agent
• CPU • Memory • Load Average • Disk Usage
• Disk I/O • Network I/O CloudWatch Datadog Agent
CloudWatch
None
ΞϓϦέʔγϣϯͷࢹ
σʔλͷૹ৴ݩ σʔλͷૹ৴ઌ ௨ઌ CloudWatch Datadog Agent library agent
• ղੳͨ͠Πϕϯτ(ࠓɺࡢɺҰिؒલɺ Ұϲ݄લ) • Ͳͷ͓٬༷͔ΒͷΠϕϯτ͕ଟ͍͔ • ղੳ༻ʹཷ·͍ͬͯΔΩϡʔͷ • λΠϜΞτͷׂ߹ •
֤ॲཧʹ͔͔͍ͬͯΔ࣌ؒ • ϨεϙϯελΠϜ • ΞϓϦέʔγϣϯͷόά • Τϥʔϩά
var dogstatsd = require('libs-dogstatsd') const stats = dogstatsd.start(); // some
method stats.tick(‘test’, 1, 1, [pid:1234]); libs-dogstatsd (A wrapper library of node-dogstatsd) ࣮ߦ࣌ؒॲཧճΛDatadogʹૹ৴Ͱ͖Δϥούʔ https://github.com/makinoy/libs-dogstatsd/
Ͳ͏ͬͯࢹ͍ͯ͠Δͷ͔ʁ
LayerΛ͚ͯDashboardΛ࡞
• Layer1: KARTE͕ੜ͖͍ͯΔ͔ • Layer2: ॏཁͳϝτϦοΫ • Layer3: ֤roleͷৄࡉϝτϦοΫ •
(Layer4: ΠϯελϯεຖͷϝτϦοΫ)
None
None
None
None
Ұݩཧ͢Δ͜ͱ υϦϧμϯͯ͠Λ͍͚ͬͯΔ͜ͱ
࣍ 1. αʔϏεͷհ 2. ࢹͷత 3. Πϯϑϥߏ 4. ࢹͷํ๏ 5.
Ξϥʔτ 6. ϋϚͬͨ͜ͱ 7. ·ͱΊ
جຊతʹΞϥʔτ DatadogͰҰݩཧ͢Δ
ϝτϦοΫ τϦΨʔ ௨ઌ MCM ᮢ ҟৗ ▪Πϯϑϥ
ϝτϦοΫ τϦΨʔ ௨ઌ ᮢ ҟৗ ▪ΞϓϦέʔγϣϯ ᮢ
ҟৗݕग़͕ศར
None
࣍ 1. αʔϏεͷհ 2. ࢹͷత 3. Πϯϑϥߏ 4. ࢹͷํ๏ 5.
Ξϥʔτ 6. ϋϚͬͨ͜ͱ 7. ·ͱΊ
1.ϞχλϦϯάπʔϧͷྉ͔͔ۚΓ͗͢
ɾΠϯελϯεͷىಈGCPʹൺ͍ͯ ɾASΛ͏ͱҰʹىಈ/ऴྃ͢ΔΛઃఆͰ͖Δ ɾΠϯελϯεͷىಈAWSʹൺͯૣ͍ ɾASͰىಈ/ऴྃ͢Δͪ͜ΒͰઃఆͰ͖ͳ͍ ɾͨ͠ϗετ͋ͨΓ$18/ ɾΛ͑Δͱ1͋ͨΓ$0.03/hr
AutoScaler͕ҰؾʹΠϯελϯεΛ૿ͯ͠ɺҰ ؾʹΠϯελϯεΛݮΒ͢ͱ͍͏ڍಈΛྑ͘͢Δ Datadog Agent͕ىಈͨ͠Πϯελϯε͕ܹ૿ GCPΛ͍࢝Ί͔ͯΒલ݄ͷྉۚͷ2ഒΛ͑Δྉ ͕ۚٻ͞ΕΔ "# αϙʔτʹ૬ஊͯ͠࠷దͳ՝ۚମܥΛఏҊͯ͠Β ͍ɺ$1000Ҏ্҆͘ͳͬͨ $
ΠϯελϯεͷಛੑͱࢹαʔϏεͷ՝ۚମܥ Λ͔ͬ͠ΓѲ͓ͯ͘͠ʂ
2. ϝτϦοΫૹ৴λΠϛϯά͕ఆͱҧ͏
Քಇ͍ͯ͠ΔΠϯελϯεͷ͕ ຖૹΒΕ͍ͯΔΑ͏ʹݟ͑Δ%
೦5ຖͰͨ͠ʂ& ΞϥʔτΛֻ͚ͯؾͮ͘ͷ͕5ޙ'
ϝτϦοΫͷૹ৴සΛѲ͓ͯ͘͠ʂ
࣍ 1. αʔϏεͷհ 2. ࢹͷత 3. Πϯϑϥߏ 4. ࢹͷํ๏ 5.
Ξϥʔτ 6. ϋϚͬͨ͜ͱ 7. ·ͱΊ
1. σʔλΞϥʔτҰݩཧ͢Δ 2. υϦϧμϯͯ͠Λ͍͚ͬͯΔΑ͏ʹ͢Δ 3. ֤ϓϥοτϑΥʔϜͷಛੑͱɺࢹπʔϧͷ՝ۚ ମܥΛѲ͢Δ 4. ΠϯςάϨʔγϣϯϝτϦοΫͷૹ৴සΛѲ ͢Δ
ࠓޙ
1. DatadogʹΞϥʔτΛ·ͱΊΔ 2. Datadog͕SPOFʹͳ͍ͬͯΔ݅Λ࠶ݕ౼͢Δ 3. ࢹઃఆͷexportΛ͍ͨ͠
ΤϯδχΞืूதͰ͢ʂ