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
La métrologie dans les applications Scala
Search
Sponsored
·
SiteGround - Reliable hosting with speed, security, and support you can count on.
→
Matthieu Guillermin
October 28, 2016
Programming
0
150
La métrologie dans les applications Scala
Talk Scala.IO 2016
Matthieu Guillermin
October 28, 2016
Tweet
Share
Other Decks in Programming
See All in Programming
Go 1.26でのsliceのメモリアロケーション最適化 / Go 1.26 リリースパーティ #go126party
mazrean
1
400
AIに任せる範囲を安全に広げるためにやっていること
fukucheee
0
130
RubyとGoでゼロから作る証券システム: 高信頼性が求められるシステムのコードの外側にある設計と運用のリアル
free_world21
0
290
PostgreSQL を使った快適な go test 環境を求めて
otakakot
0
550
AI 開発合宿を通して得た学び
niftycorp
PRO
0
120
コードレビューをしない選択 #でぃーぷらすトウキョウ
kajitack
3
960
AI駆動開発の本音 〜Claude Code並列開発で見えたエンジニアの新しい役割〜
hisuzuya
4
510
SourceGeneratorのマーカー属性問題について
htkym
0
200
AIとペアプロして処理時間を97%削減した話 #pyconshizu
kashewnuts
1
240
[SF Ruby Feb'26] The Silicon Heel
palkan
0
100
Ruby x Terminal
a_matsuda
7
600
CS教育のDX AIによる育成の効率化
niftycorp
PRO
0
120
Featured
See All Featured
Practical Orchestrator
shlominoach
191
11k
StorybookのUI Testing Handbookを読んだ
zakiyama
31
6.6k
Git: the NoSQL Database
bkeepers
PRO
432
66k
Why Your Marketing Sucks and What You Can Do About It - Sophie Logan
marketingsoph
0
110
Darren the Foodie - Storyboard
khoart
PRO
3
2.9k
What the history of the web can teach us about the future of AI
inesmontani
PRO
1
470
Digital Projects Gone Horribly Wrong (And the UX Pros Who Still Save the Day) - Dean Schuster
uxyall
0
720
The Art of Delivering Value - GDevCon NA Keynote
reverentgeek
16
1.9k
Joys of Absence: A Defence of Solitary Play
codingconduct
1
310
コードの90%をAIが書く世界で何が待っているのか / What awaits us in a world where 90% of the code is written by AI
rkaga
60
42k
How to build a perfect <img>
jonoalderson
1
5.3k
My Coaching Mixtape
mlcsv
0
71
Transcript
La métrologie dans les applications Scala Matthieu Guillermin Scala.IO 2016
Matthieu Guillermin Lead Dev Brainify - Analytics e-commerce
Est-ce que mon service fonctionne ? Est-ce que les temps
de réponse sont corrects ? Est-ce que mon service renvoie des erreurs 500 ? Il y a beaucoup d'erreurs, qu'est-ce qui cloche ? « Ça fait 1 heure que l'appli ne fonctionne plus, qu'est-ce qu'il se passe ? »
Logs • Disponible dans toutes les applications • Très détaillé
• Difficile à exploiter • Difficile d'avoir une vision synthétique
Métriques • Valeurs numériques • Facile à stocker • Facile
à exploiter • Affichage • Triggers / Alertes
Statsd / Graphite Cabot Service A Kamon Service B Kamon
Service C Kamon Grafana system C* ES Kafka Kong
Enregistrement des métriques
Kamon • JVM / Scala • Reactive-friendly • Modulaire •
http://kamon.io
Kamon - Metrics • Flush interval • Instruments • Counter
• Histogram • ...
// Counter val authCounter = Kamon.metrics.counter("auth") def auth(username: String, password:
String) = { authCounter.increment() // ... } // Histogram val docSizeHistogram = Kamon.metrics.histogram("doc-size") def storeDocument(doc: Document) = { docSizeHistogram.record(doc.size) // ... }
Kamon - Trace • Enregistrement du "temps passé" • Granularité
: Traces / Segments
def userCreate(user: User) = { val context = Kamon.tracer.newContext("userCreate") //
... validateUser(user) // ... context.finish() } def validateUser(user: User) = { val segment = Tracer.currentContext.startSegment("validateUser", "myCategory", "myLib") // ... segment.finish() }
def userCreate(user: User) = { Tracer.withNewContext("userCreate", autoFinish = true) {
validateUser(user) // ... } } def validateUser(user: User) = { Tracer.currentContext.withNewSegment("validateUser", "cat", "lib") { // ... } }
def notifyUser(user: User) = { Tracer.currentContext.withNewAsyncSegment("notifyUser", "cat", "lib") { sendEmail(user.email)
// sendEmail() is returning a Future } }
Kamon - Intégrations • JVM : • Heap, GC, ...
• Akka : Metrics sur Actor, Router, Dispatcher • Ex : time-in-mailbox, processing-time, errors,... • Play : Metrics sur requêtes HTTP • Ex : requêtes entrantes, WSClient,...
class User @Inject()(val userService: UserService) extends Controller { def create(username:
String, password: String) = TraceName("user-create") { Action.async { userService.create(username, password) .map(user => Ok(Json.toJson(user))) } } }
kamon { statsd { hostname = my.statsd.server flush-interval = 1
minute simple-metric-key-generator { application = "my-service" } } metric { tick-interval = 1 minute filters { akka-actor { includes = ["application/user/**"] # excludes = [] } } } system-metrics { sigar-enabled = false jmx-enabled = true } }
Visualisation des métriques
Grafana • Visualisation des métriques • Dashboards • Templating /
annotations • http://grafana.org/
None
None
None
Alerting
Alerting • Etre alerté d'un incident - hits / last
2 minutes = 0 - 2xx_status = 0
Alerting • Prévenir un incident - heap_usage > 90% -
disk_free < 10% - response_time > 200ms
Cabot • Services / Checks • ICMP / HTTP /
Graphite • Notifications • Slack • Twilio (SMS / Phone) http://cabotapp.com/
None
None
None
None
Conclusion • Mettre en place de la métrologie • Investissement
qui en vaut la peine • Simple • Dev + Ops • Processus itératif