🥇Panyebaran kenari otomatis nganggo Flagger lan Istio

CD diakoni minangka praktik piranti lunak perusahaan lan minangka evolusi alami saka prinsip CI sing mapan. Nanging, CD isih cukup langka, bisa uga amarga kerumitan manajemen lan wedi yen panyebaran gagal bisa mengaruhi kasedhiyan sistem.

Panji minangka operator Kubernetes open source sing tujuane kanggo ngilangi hubungan sing mbingungake. Iki ngotomatisasi promosi penyebaran kenari nggunakake offset lalu lintas Istio lan metrik Prometheus kanggo nganalisa prilaku aplikasi sajrone rollout sing dikelola.

Ing ngisor iki minangka pandhuan langkah-langkah kanggo nyetel lan nggunakake Flagger ing Google Kubernetes Engine (GKE).

Nggawe kluster Kubernetes

Sampeyan miwiti kanthi nggawe kluster GKE nganggo add-on Istio (yen sampeyan ora duwe akun GCP, sampeyan bisa ndhaptar kene - kanggo nampa kridit gratis).

Mlebu menyang Google Cloud, gawe proyek, lan aktifake tagihan. Instal utilitas baris perintah gcloud lan ngatur project karo gcloud init.

Setel proyek standar, area komputasi, lan zona (ganti PROJECT_ID kanggo proyek sampeyan):

gcloud config set project PROJECT_ID
gcloud config set compute/region us-central1
gcloud config set compute/zone us-central1-a

Aktifake layanan GKE lan gawe kluster nganggo tambahan HPA lan Istio:

gcloud services enable container.googleapis.com
K8S_VERSION=$(gcloud beta container get-server-config --format=json | jq -r '.validMasterVersions[0]')
gcloud beta container clusters create istio 
--cluster-version=${K8S_VERSION} 
--zone=us-central1-a 
--num-nodes=2 
--machine-type=n1-standard-2 
--disk-size=30 
--enable-autorepair 
--no-enable-cloud-logging 
--no-enable-cloud-monitoring 
--addons=HorizontalPodAutoscaling,Istio 
--istio-config=auth=MTLS_PERMISSIVE

Printah ing ndhuwur bakal nggawe blumbang simpul standar sing dumadi saka rong VM n1-standard-2 (vCPU: 2, RAM 7,5 GB, disk: 30 GB). Saenipun, komponen Istio kudu diisolasi saka beban kerja, nanging ora ana cara sing gampang kanggo mbukak pod Istio ing blumbang simpul khusus. Manifestasi Istio dianggep mung diwaca, lan GKE bakal mbalekake owah-owahan apa wae kayata naleni simpul utawa pisah saka pod.

Nggawe kredensial kanggo kubectl:

gcloud container clusters get-credentials istio

Nggawe ikatan peran admin cluster:

kubectl create clusterrolebinding "cluster-admin-$(whoami)" 
--clusterrole=cluster-admin 
--user="$(gcloud config get-value core/account)"

Instal alat baris perintah Helm:

brew install kubernetes-helm

Homebrew 2.0 saiki uga kasedhiya kanggo Linux.

Nggawe akun layanan lan ngiket peran kluster kanggo Tiller:

kubectl -n kube-system create sa tiller && 
kubectl create clusterrolebinding tiller-cluster-rule 
--clusterrole=cluster-admin 
--serviceaccount=kube-system:tiller

Expand Tiller ing namespace kube-system:

helm init --service-account tiller

Sampeyan kudu nimbang nggunakake SSL antarane Helm lan Tiller. Kanggo informasi luwih lengkap babagan nglindhungi instalasi Helm, waca docs.helm.sh

Konfirmasi setelan:

kubectl -n istio-system get svc

Sawise sawetara detik, GCP kudu menehi alamat IP eksternal menyang layanan kasebut istio-ingressgateway.

Nggawe Istio Ingress Gateway

Nggawe alamat IP statis kanthi jeneng istio-gatewaynggunakake alamat IP gateway Istio:

export GATEWAY_IP=$(kubectl -n istio-system get svc/istio-ingressgateway -ojson | jq -r .status.loadBalancer.ingress[0].ip)
gcloud compute addresses create istio-gateway --addresses ${GATEWAY_IP} --region us-central1

Saiki sampeyan butuh domain internet lan akses menyang registrar DNS sampeyan. Tambah rong cathetan A (ganti example.com menyang domain sampeyan):

istio.example.com   A ${GATEWAY_IP}
*.istio.example.com A ${GATEWAY_IP}

Priksa manawa wildcard DNS bisa digunakake:

watch host test.istio.example.com

Gawe gateway Istio umum kanggo nyedhiyakake layanan ing njaba jaringan layanan liwat HTTP:

apiVersion: networking.istio.io/v1alpha3
kind: Gateway
metadata:
  name: public-gateway
  namespace: istio-system
spec:
  selector:
    istio: ingressgateway
  servers:
    - port:
        number: 80
        name: http
        protocol: HTTP
      hosts:
        - "*"

Simpen sumber ing ndhuwur minangka public-gateway.yaml banjur gunakake:

kubectl apply -f ./public-gateway.yaml

Ora ana sistem produksi sing kudu nyedhiyakake layanan ing Internet tanpa SSL. Kanggo ngamanake gateway ingress Istio sampeyan karo manajer sertifikat, CloudDNS lan Ayo Encrypt, waca dokumentasi Panji G.K.E.

Pemasangan Flagger

Tambahan GKE Istio ora kalebu conto Prometheus sing ngresiki layanan telemetri Istio. Wiwit Flagger nggunakake metrik HTTP Istio kanggo nindakake analisis kenari, sampeyan kudu masang konfigurasi Prometheus ing ngisor iki, padha karo skema Istio Helm resmi.

REPO=https://raw.githubusercontent.com/stefanprodan/flagger/master
kubectl apply -f ${REPO}/artifacts/gke/istio-prometheus.yaml

Tambah repositori Flagger Helm:

helm repo add flagger [https://flagger.app](https://flagger.app/)

Expand Flagger kanggo namespace istio-systemkanthi ngaktifake kabar Slack:

helm upgrade -i flagger flagger/flagger 
--namespace=istio-system 
--set metricsServer=http://prometheus.istio-system:9090 
--set slack.url=https://hooks.slack.com/services/YOUR-WEBHOOK-ID 
--set slack.channel=general 
--set slack.user=flagger

Sampeyan bisa nginstal Flagger ing sembarang namespace anggere bisa komunikasi karo layanan Istio Prometheus ing port 9090.

Flagger duwe dashboard Grafana kanggo analisis kenari. Instal Grafana ing namespace istio-system:

helm upgrade -i flagger-grafana flagger/grafana 
--namespace=istio-system 
--set url=http://prometheus.istio-system:9090 
--set user=admin 
--set password=change-me

Mbukak Grafana liwat gateway mbukak kanthi nggawe layanan virtual (ganti example.com menyang domain sampeyan):

apiVersion: networking.istio.io/v1alpha3
kind: VirtualService
metadata:
  name: grafana
  namespace: istio-system
spec:
  hosts:
    - "grafana.istio.example.com"
  gateways:
    - public-gateway.istio-system.svc.cluster.local
  http:
    - route:
        - destination:
            host: flagger-grafana

Simpen sumber ing ndhuwur minangka grafana-virtual-service.yaml banjur gunakake:

kubectl apply -f ./grafana-virtual-service.yaml

Nalika arep http://grafana.istio.example.com Browser sampeyan kudu ngarahake sampeyan menyang kaca login Grafana.

Nyebarake aplikasi web nganggo Flagger

Flagger nyebarke Kubernetes lan, yen perlu, autoscaling horisontal (HPA), banjur nggawe seri obyek (penyebaran Kubernetes, layanan ClusterIP lan layanan virtual Istio). Objek kasebut nyedhiyakake aplikasi menyang bolong layanan lan ngatur analisis lan promosi kenari.

Gawe ruang jeneng tes kanthi implementasi Istio Sidecar diaktifake:

REPO=https://raw.githubusercontent.com/stefanprodan/flagger/master
kubectl apply -f ${REPO}/artifacts/namespaces/test.yaml

Gawe panyebaran lan alat skala horisontal otomatis kanggo pod:

kubectl apply -f ${REPO}/artifacts/canaries/deployment.yaml
kubectl apply -f ${REPO}/artifacts/canaries/hpa.yaml

Pasang layanan tes beban kanggo ngasilake lalu lintas sajrone analisis kenari:

helm upgrade -i flagger-loadtester flagger/loadtester 
--namepace=test

Gawe sumber kenari khusus (ganti example.com menyang domain sampeyan):

apiVersion: flagger.app/v1alpha3
kind: Canary
metadata:
  name: podinfo
  namespace: test
spec:
  targetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: podinfo
  progressDeadlineSeconds: 60
  autoscalerRef:
    apiVersion: autoscaling/v2beta1
    kind: HorizontalPodAutoscaler
    name: podinfo
  service:
    port: 9898
    gateways:
    - public-gateway.istio-system.svc.cluster.local
    hosts:
    - app.istio.example.com
  canaryAnalysis:
    interval: 30s
    threshold: 10
    maxWeight: 50
    stepWeight: 5
    metrics:
    - name: istio_requests_total
      threshold: 99
      interval: 30s
    - name: istio_request_duration_seconds_bucket
      threshold: 500
      interval: 30s
    webhooks:
      - name: load-test
        url: http://flagger-loadtester.test/
        timeout: 5s
        metadata:
          cmd: "hey -z 1m -q 10 -c 2 http://podinfo.test:9898/"

Simpen sumber ing ndhuwur minangka podinfo-canary.yaml banjur gunakake:

kubectl apply -f ./podinfo-canary.yaml

Analisis ing ndhuwur, yen sukses, bakal mlaku limang menit, mriksa metrik HTTP saben setengah menit. Sampeyan bisa nemtokake wektu minimal sing dibutuhake kanggo nyoba lan ningkatake penyebaran kenari kanthi nggunakake rumus ing ngisor iki: interval * (maxWeight / stepWeight). Kolom CRD Canary didokumentasikan kene.

Sawise sawetara detik, Flagger bakal nggawe obyek kenari:

# applied 
deployment.apps/podinfo
horizontalpodautoscaler.autoscaling/podinfo
canary.flagger.app/podinfo
# generated 
deployment.apps/podinfo-primary
horizontalpodautoscaler.autoscaling/podinfo-primary
service/podinfo
service/podinfo-canary
service/podinfo-primary
virtualservice.networking.istio.io/podinfo

Bukak browser lan pindhah menyang app.istio.example.com, sampeyan kudu ndeleng nomer versi aplikasi demo.

Analisis lan promosi kenari otomatis

Flagger ngetrapake loop kontrol sing mboko sithik mindhah lalu lintas menyang kenari nalika ngukur indikator kinerja utama kayata tingkat sukses panjalukan HTTP, durasi panjalukan rata-rata, lan kesehatan pod. Adhedhasar analisis KPI, kenari dipromosikan utawa dihentikan, lan asil analisis kasebut diterbitake ing Slack.

Penyebaran Canary dipicu nalika salah siji saka obyek ing ngisor iki diganti:

Deploy PodSpec (gambar wadah, printah, port, env, lsp.)
ConfigMaps dipasang minangka volume utawa diowahi dadi variabel lingkungan
Rahasia dipasang minangka volume utawa diowahi dadi variabel lingkungan

Jalanake penyebaran kenari nalika nganyari gambar wadah:

kubectl -n test set image deployment/podinfo 
podinfod=quay.io/stefanprodan/podinfo:1.4.1

Flagger ndeteksi manawa versi panyebaran wis diganti lan wiwit nganalisa:

kubectl -n test describe canary/podinfo

Events:

New revision detected podinfo.test
Scaling up podinfo.test
Waiting for podinfo.test rollout to finish: 0 of 1 updated replicas are available
Advance podinfo.test canary weight 5
Advance podinfo.test canary weight 10
Advance podinfo.test canary weight 15
Advance podinfo.test canary weight 20
Advance podinfo.test canary weight 25
Advance podinfo.test canary weight 30
Advance podinfo.test canary weight 35
Advance podinfo.test canary weight 40
Advance podinfo.test canary weight 45
Advance podinfo.test canary weight 50
Copying podinfo.test template spec to podinfo-primary.test
Waiting for podinfo-primary.test rollout to finish: 1 of 2 updated replicas are available
Promotion completed! Scaling down podinfo.test

Sajrone analisis, asil kenari bisa dipantau nggunakake Grafana:

Wigati dicathet: yen owah-owahan anyar ditrapake kanggo penyebaran sajrone analisis kenari, Flagger bakal miwiti maneh fase analisis.

Gawe dhaptar kabeh kenari ing kluster sampeyan:

watch kubectl get canaries --all-namespaces
NAMESPACE   NAME      STATUS        WEIGHT   LASTTRANSITIONTIME
test        podinfo   Progressing   15       2019-01-16T14:05:07Z
prod        frontend  Succeeded     0        2019-01-15T16:15:07Z
prod        backend   Failed        0        2019-01-14T17:05:07Z

Yen sampeyan wis ngaktifake kabar Slack, sampeyan bakal nampa pesen ing ngisor iki:

Rollback otomatis

Sajrone analisis kenari, sampeyan bisa ngasilake kesalahan HTTP 500 sintetik lan latensi respon dhuwur kanggo mriksa yen Flagger bakal mungkasi panyebaran kasebut.

Gawe pod test lan tindakake ing ngisor iki:

kubectl -n test run tester 
--image=quay.io/stefanprodan/podinfo:1.2.1 
-- ./podinfo --port=9898
kubectl -n test exec -it tester-xx-xx sh

Ngasilake kesalahan HTTP 500:

watch curl http://podinfo-canary:9898/status/500

Generasi tundha:

watch curl http://podinfo-canary:9898/delay/1

Nalika jumlah mriksa gagal tekan batesan, lalu lintas wis routed bali menyang saluran utami, kenari wis scaled menyang nul, lan penyebaran prajurit wis ditandhani minangka gagal.

Kesalahan canary lan lonjakan latensi dicathet minangka acara Kubernetes lan direkam dening Flagger ing format JSON:

kubectl -n istio-system logs deployment/flagger -f | jq .msg

Starting canary deployment for podinfo.test
Advance podinfo.test canary weight 5
Advance podinfo.test canary weight 10
Advance podinfo.test canary weight 15
Halt podinfo.test advancement success rate 69.17% < 99%
Halt podinfo.test advancement success rate 61.39% < 99%
Halt podinfo.test advancement success rate 55.06% < 99%
Halt podinfo.test advancement success rate 47.00% < 99%
Halt podinfo.test advancement success rate 37.00% < 99%
Halt podinfo.test advancement request duration 1.515s > 500ms
Halt podinfo.test advancement request duration 1.600s > 500ms
Halt podinfo.test advancement request duration 1.915s > 500ms
Halt podinfo.test advancement request duration 2.050s > 500ms
Halt podinfo.test advancement request duration 2.515s > 500ms
Rolling back podinfo.test failed checks threshold reached 10
Canary failed! Scaling down podinfo.test

Yen sampeyan wis ngaktifake kabar Slack, sampeyan bakal nampa pesen nalika tenggat wektu kanggo ngrampungake utawa tekan jumlah maksimum review sing gagal ing analisis ngluwihi:

Ing kesimpulan

Nganggo bolong layanan kaya Istio ing ndhuwur Kubernetes bakal nyedhiyakake metrik, log, lan log otomatis, nanging nggunakake beban kerja isih gumantung karo alat eksternal. Flagger duwe tujuan kanggo ngganti iki kanthi nambah kemampuan Istio pangiriman progresif.

Flagger kompatibel karo solusi CI/CD kanggo Kubernetes, lan analisis kenari bisa gampang ditambahi webhooks kanggo nindakake tes integrasi / ditampa sistem, tes mbukak utawa tes khusus liyane. Amarga Flagger minangka deklaratif lan nanggapi acara Kubernetes, bisa digunakake ing pipa GitOps bebarengan karo Tenun Fluks utawa JenkinsX. Yen sampeyan nggunakake JenkinsX, sampeyan bisa nginstal Flagger karo tambahan jx.

Flagger didhukung Tenun lan nyedhiyakake penyebaran kenari ing Tenun Awan. Proyèk iki dites ing GKE, EKS lan bare metal karo kubeadm.

Yen sampeyan duwe saran kanggo nambah Flagger, kirimake masalah utawa PR ing GitHub ing stefanprodan / flagger. Kontribusi luwih saka sambutan!

Спасибо Ray Tsang.

Source: www.habr.com

Penyebaran kenari otomatis kanthi Flagger lan Istio