Shakr - Google Container Engine을 이용한 빠르고 안정적인 서비스 개발 및 배포

Harder, better, faster, stronger with Google Container Engine ҳӖஶప੉ցূ૓ਸࢎਊೠࡅܰҊউ੿੸ੋࢲ࠺झѐߊ߂ߓನ CTO
MINKU LEE

v ⿏࠺٣য়ҟҊ Container icon by Dominic Plag from the Noun
Project ֥ۄ਎੿ب۽рױೠ࠺٣য়ઁ੘ݧ૓࠺٣য়٣੗ੋ੉ೠоٙ

Shakr Infrastructure ⿏֙ஶప੉ցӝ߈ইఃఫ୊۽੹ജ ⿏֙($1ࢎਊद੘ ⿏֙؀ࠗ࠙੄ஹೊ౴झషܻ૑੗ਗਸ($1۽੉੹   $POUBJOFS&OHJOFद೷द੘ ⿏അ੤؀ࠗ࠙੄௿ۄ਋٘ஹೊ౴੘সਸ  $POUBJOFS&OHJOFীࢲҳز઺

Topics Kubernetes Container Engine Case Studies

Kubernetes

Why Container? ⿏زੌೞ٘ਝয7.ղীࢲࠁ׮ਬোೞѱܻࣗझܳ ҙܻೡࣻ੓਺ ⿏য٣ࢲա੤അоמೠগ೒ܻா੉࣌ҳزജ҃ ⿏7.؀࠺੸਷য়ߡ೻٘۽੸׼ೠѺܻࢿઁҕ١ Container icon by Dominic
Plag from the Noun Project

Kubernetes ⿏ஶప੉ցӝ߈গ೒ܻா੉࣌ਸߓನೞҊ ഛ੢ೞҊ ҙܻ೧઱חदझమ ⿏ҳӖ੉ࢎਊೞחஶప੉ցҙܻदझమ੄  ই੉٣য৬֢ೞ਋ܳ҅थೠ೐۽ં౟ ⿏৔ܻӝস੉ইצ-JOVY'PVOEBUJPOীࢲ਍৔ೞח  $MPVE/BUJWF$PNQVUJOH'PVOEBUJPOীࢲ೐۽ ં౟ܳҙܻ

Kubernetes ⿏ஶప੉ցӝ߈গ೒ܻா੉࣌झாે݂ ⿏ࢲ࠺झ٣झழߡܻ ⿏۽٘ߖ۠य ⿏ࢸ੿߂࠺޻੿ࠁҙܻ ⿏٣झ௼ҙܻ ⿏١١

࠺तೠࣗ೐౟ਝয ⿏%PDLFS4XBSN ⿏"NB[PO&$4 ⿏.FTPTQIFSF "QBDIF.FTPT ⿏)BTIJ$PSQ/PNBE

࠺तೠࣗ೐౟ਝয؀࠺ъ੼ ⿏ஶప੉ցझாે݂ࠗఠ۽٘ߖ۠य ࢸ੿ ٣झ௼ҙܻө૑  ߹ب੄بҳহ੉೧Ѿоמ ⿏($1ղীࢲݻߣ੄௿ܼਵ۽औѱ௿۞झఠࢤࢿоמ ⿏ৈ۞௿ۄ਋٘ઁҕ੗৬߬যݫఎࢲߡܳഒ೤ೞৈ  ௿۞झఠܳҳ୷ೡࣻ੓਺ $MVTUFS'FEFSBUJPO
⿏4UBUFMFTTೠগ೒ܻா੉࣌ࡺ݅੉ইצ4UBUFGVMগ೒ܻா੉࣌بҳز ೡࣻ੓ח١ৈ۞दաܻয়ী؀ೠ؀਽

Kubernetes ҳࢿਃࣗ ⿏1PE ⿏%FQMPZNFOU ⿏4FSWJDF ⿏*OHSFTT ⿏%BFNPO4FU ⿏4FDSFU ⿏$POGJH.BQ ⿏4UBUFGVM4FU
⿏3FQMJDB4FU ⿏$SPO+PC ⿏+PC ⿏1FSTJTUFOU7PMVNF ⿏1FSTJTUFOU7PMVNF$MBJN ⿏FUDj

Case Study: Shakr ӝઓੋ೐ۄܳ$POUBJOFS&OHJOFਵ۽੉੹ೞӝ

ӝઓইఃఫ୊ Redis Task Queue Fastly CDN Original Media Files Cloud
Storage Processed Media Files Cloud Storage Storage User Uploads Cloud Storage Rendered Segments Cloud Storage CDN Interconnect Upload Processing Worker Compute Engine Multiple Instances Video Processing Worker Compute Engine Multiple Instances Image Processing Worker Compute Engine Multiple Instances Video Rendering Worker Compute Engine Multiple Instances Video Design Processing Worker Compute Engine Multiple Instances Compute Papertrail Logging Datadog Monitoring

೧ѾೞҊर঻؍੼ ⿏ਬোೞ૑ޅೠஹೊ౴੗ਗഝਊ  ೠੋझఢझӒܛ੄$16ࢎਊܫ੉ੌٸҗࠗೞ࢚కੋ׮ܲੋझఢझӒܛ੉੉ܳഝਊೡࣻহਸө ⿏ߓನ੺ର੄ߣѢ۽਑  "QQ&OHJOF੉ա)FSPLVэ਷1BB4୊ۢಞܻೞѱߓನೡࣻחহਸө ⿏ৈ۞ҳࢿਃࣗܳ߹ѐ۽ҙܻ೧ঠೣ  ೾झ୓௼ ۽Ӓझ౟ܻ߁җэ਷ҳࢿਃࣗܳಞܻೞѱࢸ੿ೡࣻחহਸө

Compute Engine ীࢲ Container Engine ਵ۽ ⿏$POUBJOFS&OHJOF௿۞झఠࢤࢿ ⿏%FQMPZNFOU4QFD੘ࢿ ⿏ߓನ߂܀݂সؘ੉౟ ⿏/PEF1PPM୶о߂࢏ઁ
⿏%BUBEPH"HFOUܳాೠݽפఠ݂ ⿏۽Ӓझ౟ܻ߁߂੷੢

Container Engine ௿۞झఠ ࢤࢿ

Google Container Engine Kubernetes as a Service

Google Container Engine ⿏,VCFSOFUFTOPEF৬FUDEੋझఢझܳ(PPHMF੉ҙܻ ⿏੗زਵ۽(PPHMF$PNQVUF&OHJOFੋझఢझࢤࢿ߂ࢸ੿ ⿏੗زਵ۽ߑച߷ࢸ੿ ⿏੗زਵ۽۽٘ߖ۠यࢸ੿

௿۞झఠࢤࢿ ⿏ਬোೞ૑ޅೠஹೊ౴੗ਗഝਊ  ೠੋझఢझӒܛ੄$16ࢎਊܫ੉ੌٸҗࠗೞ࢚కੋ׮ܲੋझఢझӒ ܛ੉੉ܳഝਊೡࣻহਸө ⿏ߓನ੺ର੄ߣѢ۽਑  "QQ&OHJOF੉ա)FSPLVэ਷1BB4୊ۢಞܻೞѱߓನೡࣻ੓ਸө ⿏ৈ۞ҳࢿਃࣗܳ߹ѐ۽ҙܻ೧ঠೣ  ۽٘ߖ۠ࢲ
೾झ୓௼ ۽Ӓझ౟ܻ߁җэ਷ҳࢿਃࣗܳಞܻೞѱࢸ੿ೡ ࣻחহਸө

௿۞झఠࢤࢿ $ gcloud container clusters create my-awesome-cluster \ --zone asia-northeast1-a
\ --enable-cloud-logging --enable-cloud-monitoring \ --machine-type n1-standard-4 \ —-num-nodes 3 Bash

$ gcloud container clusters get-credentials \ my-awesome-cluster --zone asia-northeast1-a Fetching
cluster endpoint and auth data. kubeconfig entry generated for my-awesome-cluster. $ kubectl proxy Starting to serve on 127.0.0.1:8001 Bash Kubernetes Dashboard

Kubernetes Dashboard

Deployment Spec ੘ࢿ

Pod ↟ о੢੘਷ஹೊ౴ਬ׫ ↟ ೞա഑਷ৈ۞ѐ੄ஶప੉ց۽ҳࢿ ↟ 1PEউ੄ஶప੉ցח੷੢ҕрҗஶఫझ౟ܳҕਬ ↟ ൃߊࢿઙܐदݽؘٚ੉ఠਬप ↟
೙ਃद($4ա1FSTJTUFOU7PMVNFਸࢎਊೞৈؘ੉ఠܳࠁઓ

Deployment ⿏ৈ۞ѐ੄1PEਸߓನೞҊҙܻೞחܻࣗझ ⿏1PE੄іࣻ SFQMJDB ܳ૑੿೧઱ݶ੿೧૓іࣻ݅ఀ  ੗زਵ۽୶о߂࢏ઁ ⿏܀݂সؘ੉౟৬܀ߔҳഅ 3FQMJDB4FU

yaml deployment.yml apiVersion: apps/v1beta1 kind: Deployment metadata: name: myapp-deployment

yaml deployment.yml apiVersion: apps/v1beta1 kind: Deployment metadata: name: myapp-deployment spec:
replicas: 3

replicas: 3 template: metadata: labels: app: myapp spec: containers: - name: myapp image: gcr.io/my-project/myapp:3.0.0 resources: requests: cpu: 200m memory: 100Mi limits: cpu: "4" memory: 8Gi

replicas: 3 template: metadata: labels: app: myapp spec: containers: - name: myapp image: gcr.io/my-project/myapp:3.0.0 resources: requests: cpu: 200m memory: 100Mi limits: cpu: "4" memory: 8Gi livenessProbe: httpGet: path: / port: 80 readinessProbe: httpGet: path: / port: 80 initialDelaySeconds: 5 periodSeconds: 3

replicas: 3 template: metadata: labels: app: myapp spec: containers: - name: myapp image: gcr.io/my-project/myapp:3.0.0 resources: requests: cpu: 200m memory: 100Mi limits: cpu: "4" memory: 8Gi livenessProbe: httpGet: path: / port: 80 readinessProbe: httpGet: path: / port: 80 initialDelaySeconds: 5 periodSeconds: 3 env: - name: RACK_ENV value: production - name: PORT value: 80 volumeMounts: - name: tmp mountPath: /tmp volumes: - name: tmp emptyDir: {}

Deployment ߓನೞӝ Bash # Get credentials $ gcloud container clusters
get-credentials my-awesome-cluster Fetching cluster endpoint and auth data. kubeconfig entry generated for my-awesome-cluster.

Deployment ߓನೞӝ Bash # Get credentials $ gcloud container clusters
get-credentials my-awesome-cluster Fetching cluster endpoint and auth data. kubeconfig entry generated for my-awesome-cluster. # Deploy! $ kubectl apply -f path/to/deployment.yml deployment "myapp-deployment" created

Deployment ܀݂সؘ੉౟ Bash kubectl apply -f path/to/deployment.yml deployment "myapp-deployment" updated

Node Pool୶о߂࢏ઁ

Node Pool୶о߂࢏ઁ ⿏ਬোೞ૑ޅೠஹೊ౴੗ਗഝਊ  ೠੋझఢझӒܛ੄$16ࢎਊܫ੉ੌٸҗࠗೞ࢚కੋ׮ܲੋझఢझӒ ܛ੉੉ܳഝਊೡࣻহਸө ⿏ߓನ੺ର੄ߣѢ۽਑  "QQ&OHJOF੉ա)FSPLVэ਷1BB4୊ۢಞܻೞѱߓನೡࣻ੓ਸө ⿏ৈ۞ҳࢿਃࣗܳ߹ѐ۽ҙܻ೧ঠೣ 
۽٘ߖ۠ࢲ ೾झ୓௼ ۽Ӓझ౟ܻ߁җэ਷ҳࢿਃࣗܳಞܻೞѱࢸ੿ೡ ࣻחহਸө

Node Pool୶о߂࢏ઁ $ gcloud container node-pools create my-new-nodepool \ --cluster
my-awesome-cluster \ --machine-type n1-standard-16 \ --num-nodes 5 Bash

Preemptible Node Pool

Preemptible VM (PVM)

Preemptible Node Pool $ gcloud beta container node-pools create my-new-nodepool
\ --cluster my-awesome-cluster \ --machine-type n1-standard-16 \ --num-nodes 5 \ --preemptible Bash BETA

PVM with nodeSelector apiVersion: v1 kind: Pod spec: nodeSelector: !cloud.google.com/gke-preemptible
yaml

PVM with nodeAﬃnity apiVersion: v1 kind: Pod spec: affinity: nodeAffinity:
preferredDuringSchedulingIgnoredDuringExecution: nodeSelectorTerms: - matchExpressions: - key: cloud.google.com/gke-preemptible operator: DoesNotExist yaml BETA

Datadog Agent ܳాೠݽפఠ݂

Datadog Agentࢸ஖ $ kubectl create -f dd-agent.yaml daemonset "dd-agent" created
Bash

ChatOps োز

۽Ӓझ౟ܻ߁߂੷੢

௿۞झఠࢸ੿ $ gcloud container clusters create my-awesome-cluster \ --zone asia-northeast1-a
\ --enable-cloud-logging --enable-cloud-monitoring \ --machine-type n1-standard-4 \ —-num-nodes 3 Bash

۽Ӓী੉੹౟

Stackdriver Logging ⿏$POUBJOFS&OHJOFਸॵ׮ݶਗ௿ܼࢸ੿оמ ⿏ఫझ౟+40/ನݘ૑ਗ ⿏पदр۽Ӓझ౟ܻ߁ ⿏ޖܐ౭যઁҕ ೐۽ં౟׼ೠ׳(# ੌਬ૑ ⿏"84ܻࣗझبݽפఠ݂оמ
⿏$MPVE4UPSBHF #JH2VFSZ 1VC4VCਵ۽۽ӒFYQPSUоמ

Stackdriver Logging

BigQuery Export

੉੹റ ⿏ ೠੋझఢझӒܛ੄$16ࢎਊܫ੉ੌٸҗࠗೞ࢚కੋ׮ܲੋझఢझӒܛ੉੉ܳഝਊೡࣻহਸө   ª੔ৈ੗ਗ୭ࣗച ⿏ "QQ&OHJOF੉ա)FSPLVэ਷1BB4୊ۢಞܻೞѱߓನೡࣻחহਸө   ªLVCFDUMழݔ٘ۄੋبҳ۽ए਍ߓನ੗زച ⿏
೾झ୓௼ ۽Ӓझ౟ܻ߁җэ਷ҳࢿਃࣗܳಞܻೞѱࢸ੿ೡࣻחহਸө   ª؀ࠗ࠙੄ੋ೐ۄҳࢿਃࣗܳऔѱࢸ੿оמ

Case Study: StoriesAds ࢜۽਍ࢲ࠺झܳ$POUBJOFS&OHJOFী۠டೞӝ

Make Instagram Stories Ads for your brand.

Schedule April 2017 2 9 16 23 30 3 10
17 24 4 11 18 25 5 12 19 26 6 13 20 27 7 14 21 28 8 15 22 29

Schedule April 2017 2 9 16 23 30 3 10
17 24 4 11 18 25 5 12 19 26 6 13 20 27 7 14 21 28 8 15 22 29 ൞ݎ۠டੌ

Schedule April 2017 2 9 16 23 30 3 10
17 24 4 11 18 25 5 12 19 26 6 13 20 27 7 14 21 28 8 15 22 29 ೐۽షఋ੉ೝ ೐۽ં౟ࣇস ൞ݎ۠டੌ

Schedule April 2017 2 9 16 23 30 3 10
17 24 4 11 18 25 5 12 19 26 6 13 20 27 7 14 21 28 8 15 22 29 ೐۽షఋ੉ೝ ೐۽ં౟ࣇস ൞ݎ۠டੌ ߔূ٘೐ۿ౟ূ٘ѐߊ $*ജ҃ࣇস

Schedule April 2017 2 9 16 23 30 3 10
17 24 4 11 18 25 5 12 19 26 6 13 20 27 7 14 21 28 8 15 22 29 ೐۽షఋ੉ೝ ೐۽ં౟ࣇস ൞ݎ۠டੌ ߔূ٘೐ۿ౟ূ٘ѐߊ $*ജ҃ࣇস ೐۽؋࣌ജ҃ҳ୷

Challenges ⿏ઁೠػੌ੿  ࢜۽਍بҳٜܳ݅ӝਤ೧ӝഥ࠺ਊਸऔࢎܻ૑୹ೡࣻহח࢚ട ⿏੹ࣁ҅ࢎਊ੗ܳ؀࢚ਵ۽ೠࢲ࠺झ  %BZࠗఠ੹ࣁ҅ࢎਊ੗ীѱ௝੸ೠࣘبܳઁҕ೧઱যঠೠ׮ ⿏৘ஏೞӝ൨ٚߑޙ੗ࣻ  ੸ѱࢎਊೡٸח҃ઁ੸ਵ۽਍৔ೞҊ ࢎਊ੗оૐоೣীٮۄਬোೞѱഛ੢ೡࣻ੓যঠೠ׮

Challenges Container Engine

ইఃఫ୊ Container Engine Redis Task Queue Stackdriver Logging Cloud CDN
Container Builder Docker image builds Frontend k8s Deployment Multiple Pods Backend k8s Deployment Multiple Pods Worker k8s Deployment Multiple Pods Datadog Monitoring Cloud Datastore Frontend k8s Service Backend k8s Service StoriesAds k8s Ingress GitHub Source Repository Travis CI Continuous Integration Cloud Load Balancing

Container Builder Container Engine Redis Task Queue Stackdriver Logging Cloud
CDN Container Builder Docker image builds Frontend k8s Deployment Multiple Pods Backend k8s Deployment Multiple Pods Worker k8s Deployment Multiple Pods Datadog Monitoring Cloud Datastore Frontend k8s Service Backend k8s Service StoriesAds k8s Ingress GitHub Source Repository Travis CI Continuous Integration Cloud Load Balancing

Container Builder ⿏$MPVE4PVSDF3FQPTJUPSZ (JU)VC #JUCVDLFU૑ਗ ⿏࠳ے஖աకӒоੌ੿ઑѤীݏਵݶ࠽٘प೯ ⿏۽Ӓपदрഛੋоמ ⿏ޖܐ౭যઁҕ ೐۽ં౟׼ೞܖ࠙࠽٘दрઁҕ
⿏ৈ۞ஶప੉ց߽ܳ۳۽ҳزоמ ⿏ݒפ૑٘ࢲ࠺झ

yaml cloudbuild.yaml steps: - name: 'gcr.io/keynote-demo-project/node:7.9.0' args: ['yarn'] id: 'install-packages'

- name: 'gcr.io/keynote-demo-project/node:7.9.0' args: ['yarn', 'run', 'build'] env: - 'BACKEND=staging' - 'COMMIT_SHA=$COMMIT_SHA' waitFor: ['install-packages'] id: 'staging' - name: 'gcr.io/keynote-demo-project/node:7.9.0' args: ['yarn', 'run', 'build'] env: - 'BACKEND=production' - 'COMMIT_SHA=$COMMIT_SHA' waitFor: ['install-packages'] id: 'production'

- name: 'gcr.io/keynote-demo-project/node:7.9.0' args: ['yarn', 'run', 'build'] env: - 'BACKEND=staging' - 'COMMIT_SHA=$COMMIT_SHA' waitFor: ['install-packages'] id: 'staging' - name: 'gcr.io/keynote-demo-project/node:7.9.0' args: ['yarn', 'run', 'build'] env: - 'BACKEND=production' - 'COMMIT_SHA=$COMMIT_SHA' waitFor: ['install-packages'] id: 'production' - name: 'gcr.io/cloud-builders/docker' args: [ 'build', '-t', 'gcr.io/$PROJECT_ID/frontend:staging-$COMMIT_SHA', '--build-arg', 'backend=staging', '.' ] waitFor: ['staging'] - name: 'gcr.io/cloud-builders/docker' args: [ 'build', '-t', 'gcr.io/$PROJECT_ID/frontend:production-$COMMIT_SHA', '--build-arg', 'backend=production', '.' ] waitFor: ['production']

- name: 'gcr.io/keynote-demo-project/node:7.9.0' args: ['yarn', 'run', 'build'] env: - 'BACKEND=staging' - 'COMMIT_SHA=$COMMIT_SHA' waitFor: ['install-packages'] id: 'staging' - name: 'gcr.io/keynote-demo-project/node:7.9.0' args: ['yarn', 'run', 'build'] env: - 'BACKEND=production' - 'COMMIT_SHA=$COMMIT_SHA' waitFor: ['install-packages'] id: 'production' - name: 'gcr.io/cloud-builders/docker' args: [ 'build', '-t', 'gcr.io/$PROJECT_ID/frontend:staging-$COMMIT_SHA', '--build-arg', 'backend=staging', '.' ] waitFor: ['staging'] - name: 'gcr.io/cloud-builders/docker' args: [ 'build', '-t', 'gcr.io/$PROJECT_ID/frontend:production-$COMMIT_SHA', '--build-arg', 'backend=production', '.' ] waitFor: ['production'] images: [ 'gcr.io/$PROJECT_ID/frontend:staging-$COMMIT_SHA', 'gcr.io/$PROJECT_ID/frontend:production-$COMMIT_SHA' ]

Deployment Container Engine Redis Task Queue Stackdriver Logging Cloud CDN

yaml deployment-frontend.yaml apiVersion: apps/v1beta1 kind: Deployment metadata: name: frontend-deployment spec:
replicas: 2 template: metadata: labels: app: frontend spec: containers: - name: frontend image: gcr.io/project/frontend:staging-v01 resources: requests: memory: 100Mi ports: - containerPort: 80

yaml deployment-api.yaml apiVersion: apps/v1beta1 kind: Deployment metadata: name: api-deployment spec:
replicas: 3 template: metadata: labels: app: api spec: containers: - name: api image: gcr.io/project/api:v01 resources: requests: memory: 100Mi ports: - containerPort: 80 env: - name: RAILS_ENV value: production - name: RAILS_LOG_TO_STDOUT value: "true" - name: CONFIG_KEY valueFrom: configMapKeyRef: name: my-configs key: CONFIG_KEY - name: SECRET_KEY valueFrom: secretKeyRef: name: my-secrets key: SECRET_KEY

secret.yaml apiVersion: v1 kind: Secret metadata: name: my-secrets type: Opaque
data: SECRET_KEY: <base64_string> conﬁg_map.yaml apiVersion: v1 kind: ConfigMap metadata: name: my-configs type: Opaque data: CONFIG_KEY: <base64_string>

Container Engine Redis Task Queue Stackdriver Logging Cloud CDN Container
Builder Docker image builds Frontend k8s Deployment Multiple Pods Backend k8s Deployment Multiple Pods Worker k8s Deployment Multiple Pods Datadog Monitoring Cloud Datastore Frontend k8s Service Backend k8s Service StoriesAds k8s Ingress GitHub Source Repository Travis CI Continuous Integration Cloud Load Balancing Cloud Datastore

0 6 / 0 7/ 20 1 7, 1 4
*4 5 P ag e 1 of 1 https:/ /cloud.google.com/storage-options/

Cloud Datastore ⿏#JH5BCMFীӝ߈ਸك/P42-ؘ੉ఠ߬੉झ ⿏ࢥ٬ TIBSEJOH җࠂઁ SFQMJDBUJPO ਸ੗زਵ۽ҙܻೞৈਬোೠഛ੢оמ ⿏ੋؙझ૑ਗ ੗زࢤࢿ߂ࣻزࢤࢿ
⿏ੋझఢझ׼җӘ੉ইצ੷੢۝җয়ಌۨ੉࣌׼җӘ ⿏ޖܐ౭যઁҕ ⿏/JBOUJD 1PL䟖NPO(Pѐߊࢎ ীࢲࢎਊ઺

https:/ /cloudplatform.googleblog.com/2016/09/bringing-Pokemon-GO-to-life-on-Google-Cloud.html

Service Container Engine Redis Task Queue Stackdriver Logging Cloud CDN

Service ⿏-BZFS5$16%1-PBE#BMBODJOH ⿏ౠ੿1PEٜਸ,VCFSOFUFT௿۞झఠղীࢲ  যڌѱ੽Ӕ೧ঠೞח૑੿੄ ⿏,VCFSOFUFTղ੢4FSWJDF%JTDPWFSZ۽  1PEղীࢲ4FSWJDF੽Ӕоמ

Service Discovery $ kubectl exec -ti debug-pod -- bash root@debug-pod:/#
curl -I http://api-service HTTP/1.1 200 OK X-Frame-Options: SAMEORIGIN X-XSS-Protection: 1; mode=block X-Content-Type-Options: nosniff Content-Type: text/html X-Request-Id: 7b194765-ixz8-1qke-9572-0fa67509e500 X-Runtime: 0.001207 Bash

service-frontend.yaml apiVersion: v1 kind: Service metadata: name: frontend-service labels: app:
frontend spec: ports: - port: 80 protocol: TCP name: http selector: app: frontend type: NodePort service-api.yaml apiVersion: v1 kind: Service metadata: name: api-service labels: app: api spec: ports: - port: 80 protocol: TCP name: http selector: app: api type: NodePort

Ingress Container Engine Redis Task Queue Stackdriver Logging Cloud CDN

Ingress ⿏-BZFS)551 4 -PBE#BMBODJOH ⿏ഐझ౟੉ܴ߂҃۽ীٮۄౠ੿ߔূ٘۽ਃ୒ۄ਋౴оמ ⿏5-45FSNJOBUJPO૑ਗ ⿏(,&ীࢲח$MPVE)551 4 -PBE#BMBODFSܳ੗زਵ۽ࢤࢿ ⿏ఋࢎ-#৬׮ܰѱQSFXBSNJOHࠛ೙ਃ
⿏ਗ௿ܼਵ۽$MPVE$%/ࢸ੿оמ BETA

yaml ingress.yml apiVersion: extensions/v1beta1 kind: Ingress metadata: name: storiesads-com-ingress annotations:
kubernetes.io/ingress.global-static-ip-name: "storiesads-com" spec: tls: - secretName: tls-certificate rules: - http: paths: - path: /api/* backend: serviceName: api-service servicePort: 80 - path: /* backend: serviceName: frontend-service servicePort: 80

Cloud Load Balancing & CDN Container Engine Redis Task Queue
Stackdriver Logging Cloud CDN Container Builder Docker image builds Frontend k8s Deployment Multiple Pods Backend k8s Deployment Multiple Pods Worker k8s Deployment Multiple Pods Datadog Monitoring Cloud Datastore Frontend k8s Service Backend k8s Service StoriesAds k8s Ingress GitHub Source Repository Travis CI Continuous Integration Cloud Load Balancing

ইఃఫ୊ Container Engine Redis Task Queue Stackdriver Logging Cloud CDN

റӝ ⿏ಞܻೠ؀दࠁ٘৬LVCFDUM$-*بҳܳࢎਊೞৈݒ਋ࡅܲߓನоמ ⿏և਷ߧਤ੄दաܻয়ী؀਽оמೠ׮নೠܻࣗझઓ੤ ⿏(PPHMF$MPVE1MBUGPSNղ׮ܻܲࣗझ৬ো҅оਊ੉

ਬ੄ೡ੼ ⿏೐۽؋࣌ജ҃ীӂ੢غחੌܻࠗࣗझоই૒("оইצ߬ఋ࢚క ⿏"84١ৈ۞௿ۄ਋٘ܳ৊যݣ౭௿ۄ਋٘௿۞झఠܳ਍৔ೞҊ੗ೡٸח  ૒੽FUDE৬NBTUFS XPSLFS֢٘ܳࢸ੿೧ঠೣ ⿏.JDSPTPGU"[VSFח,VCFSOFUFT֎੉౭࠳૑ਗ ⿏$PSF04৬э਷স୓ী࠺ਊਸ૑ࠛೞҊ"84ա߬যݫఎীࢲݒפ૑٘۽਍৔оמ ⿏ӏݽоݒ਋੘਷ࢲ࠺झ੄҃਋  ࢚઱ݫݽܻয়ߡ೻٘١ਸҊ۰೮ਸٸࠗ੸੺ೡࣻ੓਺

We're Hiring! careers.shakr.com ࠺٣য়݃ா౴੄޷ېܳэ੉ٜ݅যоप࠙ਸ଺णפ׮

Thank you! Any questions? MINKU LEE with Shakr dev team

Shakr - Google Container Engine을 이용한 빠르고 안정적인 서비스 개발 및 배포

Shakr - Google Container Engine을 이용한 빠르고 안정적인 서비스 개발 및 배포

More Decks by Minku Lee

Other Decks in Technology

Featured

Transcript