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
Ahoi! Complex applications on Kubernetes
Search
Sponsored
·
Your Podcast. Everywhere. Effortlessly.
Share. Educate. Inspire. Entertain. You do you. We'll handle the rest.
→
Lukas Rieder
June 13, 2019
Technology
120
0
Share
Embed
Copy iframe code
Copy JS code
Copy link
Start on current slide
Ahoi! Complex applications on Kubernetes
This is a talk I gave at the 60th Berlin Elixir UG on Jun 13
Lukas Rieder
June 13, 2019
Transcript
69 slides
Ahoi!
Your captain today is Lukas Rieder.
I run a data consultancy. https://www.hibase.co/
I give talks. https://speakerdeck.com/Overbryd
Complex applications.
What are complex applications?
We start out simple…
We extend…
and store…
and integrate.
The world is complex. Simply manage it.
Complex applications.
Introducing Kubernetes.
Imagine an Elixir Supervisor. children = [ %{ id: MyApplication,
start: {MyApplication, :start_link, [[:hello]]} } ] {:ok, pid} = Supervisor.start_link(children, strategy: :one_for_one)
Imagine an Elixir Supervisor. children = [ %{ id: MyApplication,
start: {MyApplication, :start_link, [[:hello]]} } ] {:ok, pid} = Supervisor.start_link(children, strategy: :one_for_one)
Imagine an Elixir Supervisor. children = [ %{ id: MyApplication,
start: {MyApplication, :start_link, [[:hello]]} } ] {:ok, pid} = Supervisor.start_link(children, strategy: :one_for_one)
Imagine an Elixir Supervisor. Supervisor.count_children(pid) #=> %{active: 1, specs: 1,
supervisors: 0, workers: 1}
A Deployment in Kubernetes. apiVersion: apps/v1 kind: Deployment metadata: name:
my-app labels: app: my-app spec: # ... snip ... template: spec: containers: - name: my-container image: my-app:1.7.9 command: ["myapp", "start_link"] args: ["hello"]
Elixir Declarative programming. apiVersion: apps/v1 kind: Deployment metadata: name: my-app
labels: app: my-app spec: # ... snip ... template: spec: containers: - name: my-container image: my-app:1.7.9 command: ["myapp", "start_link"] args: ["hello"] children = [ %{ id: MyApplication, start: {MyApplication, :start_link, [[:hello]]} } ] {:ok, pid} = Supervisor.start_link(children, strategy: :one_for_one) Kubernetes
Single-responsibility principle. Job Service Ingress Deployment Supervisor Processes Elixir Kubernetes
Organized in graphs. Elixir Kubernetes
State converges.
Extend Kubernetes with Elixir.
Meet the operator pattern.
Meet the operator pattern. Operators are a combination of K8
resources.
Meet the operator pattern. It allows developers to encode domain
knowledge into an extension of the K8s API.
Custom Resource Definition Controller API Custom Resource Meet the operator
pattern.
Custom Resource Definition Controller API Custom Resource The CRD defines
your object through a schema.
Custom Resource Definition Controller API Custom Resource Resources are the
common declarative language, to describe a desired state.
Custom Resource Definition Controller API Custom Resource apiVersion: apiextensions.k8s.io/v1beta1 kind:
CustomResourceDefinition metadata: name: crontabs.stable.example.com spec: group: stable.example.com versions: - name: v1 served: true storage: true scope: Namespaced names: plural: crontabs singular: crontab kind: CronTab shortNames: - ct
Custom Resource Definition Controller API Custom Resource The API handles
CRUD, storage, changes, etc.
Custom Resource Definition Controller API Custom Resource The controller implements
domain specific knowledge.
Custom Resource Definition Controller API Custom Resource The controller manages
K8s and Custom resources for the given application domain.
Why Elixir?
Excellent state handling. Process Process
Pattern matching. Deconstruction. %{ "apiVersion" => "apps/v1", "kind" => "Deployment",
"metadata" => %{ "labels" => %{"app" => "nginx"}, "name" => "nginx-deployment", "namespace" => "default" }, "spec" => %{ "replicas" => 3, "selector" => %{"matchLabels" => %{"app" => "nginx"}}, "template" => %{ "metadata" => %{"labels" => %{"app" => "nginx"}}, "spec" => %{ "containers" => [ %{ "image" => "nginx:1.7.9", "name" => "nginx", "ports" => [%{"containerPort" => 80}] } ] } } } }
There are great libraries. https://github.com/coryodaniel/bonny https://github.com/coryodaniel/k8s
Inter-cluster communication. Solved. BeamVM BeamVM BeamVM BeamVM
Long running. High self sufficiency.
Real world examples.
hs-preview "As a customer, I want to see my campaigns
live, to check if they meet my expectations."
hs-preview www.example.com staging.example.com preview-aefd23.example.com preview-b32bef.example.com git:master git:live git:feat/abc git:feat/xyz
hs-preview Controller Ingress Deployment Service www.example.com git:feat/xyz
hs-preview Controller Ingress Deployment Service www.example.com git:feat/xyz
hs-preview Controller Ingress Deployment Service www.example.com preview-b32bef.example.com git:feat/xyz
hs-preview Ingress Deployment Service www.example.com Ingress preview-b32bef.example.com Controller git:feat/xyz
hs-preview Ingress Deployment Service www.example.com Ingress Deployment preview-b32bef.example.com Controller git:feat/xyz
hs-preview Ingress Deployment Service www.example.com Ingress Deployment Service preview-b32bef.example.com Controller
git:feat/xyz
hs-preview Ingress Deployment Service www.example.com Ingress Deployment Service preview-b32bef.example.com Domain
Controller git:feat/xyz
hs-preview preview-b32bef.example.com git:feat/xyz
hs-preview https://tech.highsnobiety.com/optimizing-our-deployment- pipeline-part-1-3fb04e727013
Software multitenancy "As a customer, I want my workflows to
run in a predictable and reliable manner."
Customer apiVersion: "internal.example.com/v1" kind: Customer metadata: name: big-client-0815 spec: features:
database: enabled: true storage: 256gb scheduler: enabled: true workers: 4 webhooks: enabled: false Software multitenancy
Customer Database Scheduler Job (k8s) Namespace (k8s) Deployment (k8s) Customer
Controller API Database Controller Scheduler Controller Software multitenancy
Software multitenancy Customer Database Scheduler Job (k8s) Namespace (k8s) Deployment
(k8s) Customer Database Scheduler Job (k8s) Namespace (k8s) Deployment (k8s) Customer Database Scheduler Job (k8s) Namespace (k8s) Deployment (k8s)
Robots building robots.
Getting started
Getting started ✓ Docker / Docker for Mac ✓ minikube
/ Kubernetes ✓ Elixir, Erlang ✓ bonni / k8s
Docker for mac is a one-stop-shop.
Example: global connection pools "As a customer, I want my
database to be running at all times, in order to make money."
Example: global connection pools App
Example: global connection pools App Worker 1
Example: global connection pools App Worker 1 Worker 2
Example: global connection pools App Worker 1 Worker 2
Example: global connection pools App Worker 1 Worker 2 pgbouncer
Example: global connection pools App Worker 1 Worker 2 pgbouncer
pgbouncer Customer A Customer B
Example: global connection pools Terminal time
Thank you for having me.
Contact:
[email protected]
More Decks by Lukas Rieder
See All by Lukas Rieder
Modeling Document Databases
overbryd
1
190
Flow Based Programming in 2 minutes
overbryd
0
120
Metabase UG, November 2018
overbryd
0
180
Flow Based Programming in Elixir
overbryd
0
220
Configuration in Elixir
overbryd
0
220
Metabase, three good practises for a hosted setup
overbryd
0
150
Elixir Deployment
overbryd
1
120
NIFs and C-Nodes
overbryd
0
120
SQL Workshop Part 1
overbryd
0
210
Other Decks in Technology
See All in Technology
Test Impact Analysis を武器に立ち向かう、AI 時代のテスト増加に対するヌーラボのアプローチ / getting-started-datadog-test-impact-analysis
k6s4i53rx
3
360
Oracle Base Database Service 技術詳細
oracle4engineer
PRO
17
120k
マネジメントされる側の仕事は「文句」を言うこと
riyaamemiya
7
6k
人の努力に頼らないセキュリティをエンジニアリングする在り方 / An approach to engineering security that does not rely on human effort
carta_engineering
0
240
異夢同船〜異なる夢でも同じ船でなんとかする〜
fortegp05
1
590
OWASP AI Agentic Top10に準拠するクラウドネイティブなガードレールの実践(CNDW 2026)
keitah
1
440
AWS Client VPNの設定でハマったポイントとその過程での気づき
amarelo_n24
1
430
パスワードのない鍵で、 データを手元に置く
ginjihayashi
0
400
共通基盤の実装に Kotlin - Quarkus を選定した理由と採用したアーキテクチャ
nealle
1
160
AWS Well-Arhitected Frameworkセキュリティの柱 について
miu_crescent
PRO
2
210
Go でトランザクション、どう書いてますか?
knwoop
1
560
認知負荷の減らし方 ~開発にも日常業務にも活かせるノウハウ~
minorun365
PRO
16
10k
Featured
See All Featured
Fireside Chat
paigeccino
43
4.1k
We Analyzed 250 Million AI Search Results: Here's What I Found
joshbly
2
2.1k
Everyday Curiosity
cassininazir
0
420
Designing Experiences People Love
moore
143
24k
Why Our Code Smells
bkeepers
PRO
340
58k
Measuring & Analyzing Core Web Vitals
bluesmoon
9
1.1k
Cheating the UX When There Is Nothing More to Optimize - PixelPioneers
stephaniewalter
287
14k
Visualizing Your Data: Incorporating Mongo into Loggly Infrastructure
mongodb
50
10k
Helping Users Find Their Own Way: Creating Modern Search Experiences
danielanewman
31
3.4k
How People are Using Generative and Agentic AI to Supercharge Their Products, Projects, Services and Value Streams Today
helenjbeal
2
400
The untapped power of vector embeddings
frankvandijk
2
2k
Build The Right Thing And Hit Your Dates
maggiecrowley
39
3.5k