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
AbemaTVでネイティブエンジニア からサーバーエンジニアに挑戦した話
Search
nghialv
March 13, 2017
Technology
3
2.8k
AbemaTVでネイティブエンジニア からサーバーエンジニアに挑戦した話
Battle Conference U30
http://bcu30.jp/
nghialv
March 13, 2017
Tweet
Share
More Decks by nghialv
See All by nghialv
Presentation for CA.go
nghialv
0
83
A consistent delivery process with GitOps style for any application on any platform
nghialv
0
460
How it works - 1.1 - What happens when you run kubectl apply command
nghialv
0
950
Why and how we build a unified CD system
nghialv
0
310
PipeCD at CyberAgent
nghialv
2
1.1k
Introdution_to_PipeCD.pdf
nghialv
2
680
The Journey of Software Delivery
nghialv
4
400
Monitoring at AbemaTV
nghialv
18
11k
ReactiveCocoa
nghialv
0
800
Other Decks in Technology
See All in Technology
製造業の課題解決に向けた機械学習の活用と、製造業特化LLM開発への挑戦
knt44kw
0
150
With Devin -AIの自律とメンバーの自立
kotanin0
2
1.1k
Mambaで物体検出 完全に理解した
shirarei24
2
200
Tableau API連携の罠!?脱スプシを夢見たはずが、逆に依存を深めた話
cuebic9bic
2
190
Microsoft Learn MCP/Fabric データエージェント/Fabric MCP/Copilot Studio-簡単・便利なAIエージェント作ってみた -"Building Simple and Powerful AI Agents with Microsoft Learn MCP, Fabric Data Agent, Fabric MCP, and Copilot Studio"-
reireireijinjin6
1
230
Vision Language Modelと自動運転AIの最前線_20250730
yuyamaguchi
3
1.1k
「育てる」サーバーレス 〜チーム開発研修で学んだ、小さく始めて大きく拡張するAWS設計〜
yu_kod
1
240
SRE新規立ち上げ! Hubbleインフラのこれまでと展望
katsuya0515
0
140
JAWS AI/ML #30 AI コーディング IDE "Kiro" を触ってみよう
inariku
3
240
バクラクによるコーポレート業務の自動運転 #BetAIDay
layerx
PRO
1
800
LLM開発を支えるエヌビディアの生成AIエコシステム
acceleratedmu3n
0
370
猫でもわかるQ_CLI(CDK開発編)+ちょっとだけKiro
kentapapa
0
3.4k
Featured
See All Featured
Evolution of real-time – Irina Nazarova, EuRuKo, 2024
irinanazarova
8
870
Building an army of robots
kneath
306
45k
I Don’t Have Time: Getting Over the Fear to Launch Your Podcast
jcasabona
33
2.4k
Sharpening the Axe: The Primacy of Toolmaking
bcantrill
44
2.4k
Let's Do A Bunch of Simple Stuff to Make Websites Faster
chriscoyier
507
140k
Keith and Marios Guide to Fast Websites
keithpitt
411
22k
Building Flexible Design Systems
yeseniaperezcruz
328
39k
The Straight Up "How To Draw Better" Workshop
denniskardys
235
140k
Making the Leap to Tech Lead
cromwellryan
134
9.5k
GraphQLとの向き合い方2022年版
quramy
49
14k
Visualizing Your Data: Incorporating Mongo into Loggly Infrastructure
mongodb
47
9.6k
Producing Creativity
orderedlist
PRO
346
40k
Transcript
AbemaTVͰωΠςΟϒΤϯδχΞ ͔ΒαʔόʔΤϯδχΞʹઓͨ͠ CyberAgent, Inc. AbemaTV։ൃہ αʔόʔαΠυΤϯδχΞ LE VAN NGHIA (ΪΞ)
• nghialv2607 • nghialv About me • 2014/11: CyberAgent 15৽ଔೖࣾ
• ~ 2015/6: AmebaΞϓϦ - iOS • ~ 2016/7: FRESH!ੜ์ૹ - iOS • ~ ࠓ : AbemaTV - Server Side ID ڵຯ • Φʔϓϯιʔε։ൃ ϓϩδΣΫτܦྺ • ΪΞ • ؔܕϓϩάϥϛϯά Hakuba, Future, MaterialKit…
”20ͷ൴Β͕ͲΜͳ͜ͱʹʮ ઓʯ͍ͯ͠Δͷ͔” ຊͷTalk BattleͷςʔϚ
ωΠςΟϒΤϯδχΞ αʔόʔΤϯδχΞ ➡
Ҡಈͷཧ༝ • ࣝΛ෯͍ͨ͘͠ • ΫϥυɺωοτϫʔΫ • େྔσʔλॲཧ • ࢄγεςϜɺϚΠΫϩαʔϏε
AbemaTVαʔόʔͷϞχλϦϯάγεςϜΛ࡞ ॳΊͯͷλεΫ
1 repository 35 repositories native server AbemaTVͷίʔυ
AbemaTVͷαʔόαΠυ Cloud DNS Cloud Load Balancer Cloud Router Cloud CDN
Pub/Sub BigQuery Cloud Storage Transcoder Kubernetes Wowza HLS MPEG-DASH DRM gRPC Microservices Protocol Buffers Varnish Alpine Linux Terraform Packer MongoDB Redis Cluster Elasticsearch Firebase
• ຊ • ϓϩδΣΫτͷࢿྉ • ݸਓGoogle CloudͷΞΧϯτ࡞ɼ ৭ʑ༡ΜͰΈΔ • AbemaTVίʔυΛಡΉ
• etcd, kubernetes…ͷίʔυΛಡΉ
GCP + Kubernetes + gRPC Google Cloud Platform Kubernetes Service
Service Service gRPC gRPC
AbemaTVͷΞʔΩςΫνϟ streaming-server gateway Media User Comment Share Question Purchase Firebase
Watchman Transcoder Ad system encoder Live Transcoder varnish api access live-gateway Compute Engine MongoDB Redis Cluster Container Engine Cloud Storage Firebase media access live admin access stateful stateless
Ϩϕϧ ྫ ΞϓϦέʔγϣϯ http, grpcͷΤϥʔɼϨΠςϯγɼ෦ঢ়ଶ ίϯςφ ίϯςφϦιʔεঢ়ଶ ΦʔέετϨʔγϣϯ ΫϥελʔͷϦιʔεɼεέδϡʔϦϯά ϗετ
(OS, Hardware) ϗετͷϦιʔεɼhardwareΤϥʔ ωοτϫʔΫ routers, switches ऩू͢Δσʔλ
࣌ܥྻDBͷબ • DalmatinerDB • InfluxDB • Prometheus • OpenTSDB •
Graphite • … “top10 time series databases”ͷϒϩάʹΑΓ
• PrometheusKubernetesͱ࿈ܞ͍ͯ͠Δ • αʔϏεσΟεΧόϦʔ: ΫϥυͷΑ͏ͳμΠφϛοΫͳڥʹendpoints͕มߋ͞ΕΔ • ίϯςφ, ΦʔέετϨʔγϣϯ ͷϝτϦΫεΛऔಘ͍͢͠ •
Prometheusͷੑೳ • ύϑΥʔϚϯεɿ • ҰͭͷPrometheusαʔό800K samples/s • Federation • ѹॖ: 12.3ഒ (ѹॖͳ͠ɿ16 bytes/sample, Prometheus: 1.3 bytes/sample) • PrometheusGrafanaͱ࿈ܞ͍ͯ͠Δ • PrometheusGolangͰॻ͔Ε͍ͯΔ PrometheusΛબͨ͠ཧ༝
AbemaTVʹ͓͚ΔPrometheus
Ռ
None
None
AlertManager Slack࿈ܞ͢ΔBot - ࣾ֎ϞόΠϧͰίϚϯυͰ γεςϜͷϝτϦΫε͕ݟ͑Δ - Τϥʔ, latency percentileͳͲ ͕ߴ͘ͳΔͱ௨͕དྷΔ
• ωΠςΟϒΤϯδχΞ͔ΒαʔόʔΤϯδχΞʹઓͷ • AbemaTVͷϞχλϦϯάγεςϜͷৄ͍͠ใ • “AbemaTVʹPrometheusͱ͍͏ϞχλϦϯάγεςϜΛಋೖͨ͠”ͷϒϩάϔ • αʔόʔΓ͍ͨωΠςΟϒΤϯδχΞ • ݴޠ͡Όͳͯ͘linux,
network, cloudؔ͢Δࣝͷ४උΛΦεεϝ • ·ͩ20ͳͷͰ৭ʑઓ͠·͠ΐ͏ʂ ·ͱΊ