implement-gitops-workflow
정보
이 스킬은 Argo CD 또는 Flux를 사용하여 쿠버네티스에 GitOps 워크플로우를 구현하며, 앱 오브 앱스 패턴, 자동 동기화, 드리프트 감지 기능을 제공합니다. Git에서 선언적으로 배포를 관리하여 다중 환경 프로모션과 자동 조정이 가능합니다. 명령형 `kubectl` 명령어에서 마이그레이션하고, 코드 리뷰 게이트를 강화하며, Git 기반 배포를 통해 감사 요구사항을 충족하는 데 활용하세요.
빠른 설치
Claude Code
추천npx skills add pjt222/agent-almanac -a claude-code/plugin add https://github.com/pjt222/agent-almanacgit clone https://github.com/pjt222/agent-almanac.git ~/.claude/skills/implement-gitops-workflowClaude Code에서 이 명령을 복사하여 붙여넣어 스킬을 설치하세요
문서
Implement GitOps Workflow
Deploy + manage K8s apps via GitOps: Argo CD / Flux → automated, auditable, repeatable.
Use When
- Declarative infra + app mgmt
- Migrate kubectl/helm → Git-driven
- Multi-env promotion (dev → staging → prod)
- Enforce code review + approval gates
- Compliance + audit via Git history
- DR w/ Git as SSOT
In
- Required: K8s cluster admin (EKS, GKE, AKS, self-hosted)
- Required: Git repo for manifests + Helm charts
- Required: Argo CD or Flux CLI
- Optional: Sealed Secrets / External Secrets Operator
- Optional: Image Updater for auto promotion
- Optional: Prometheus for sync monitoring
Do
See Extended Examples for complete configuration files and templates.
Step 1: Install Argo CD + configure repo
# Create namespace
kubectl create namespace argocd
# Install Argo CD
kubectl apply -n argocd -f https://raw.githubusercontent.com/argoproj/argo-cd/stable/manifests/install.yaml
# Wait for pods to be ready
kubectl wait --for=condition=ready pod -l app.kubernetes.io/name=argocd-server -n argocd --timeout=300s
# Install Argo CD CLI
curl -sSL -o argocd-linux-amd64 https://github.com/argoproj/argo-cd/releases/latest/download/argocd-linux-amd64
sudo install -m 555 argocd-linux-amd64 /usr/local/bin/argocd
rm argocd-linux-amd64
# Port-forward to access UI
kubectl port-forward svc/argocd-server -n argocd 8080:443 &
# Get initial admin password
ARGOCD_PASSWORD=$(kubectl -n argocd get secret argocd-initial-admin-secret -o jsonpath="{.data.password}" | base64 -d)
echo "Argo CD Admin Password: $ARGOCD_PASSWORD"
# Login via CLI
argocd login localhost:8080 --username admin --password "$ARGOCD_PASSWORD" --insecure
# Change admin password
argocd account update-password
# Add Git repository (HTTPS with token)
argocd repo add https://github.com/USERNAME/gitops-repo \
--username USERNAME \
--password "$GITHUB_TOKEN" \
--name gitops-repo
# Or add via SSH
ssh-keygen -t ed25519 -C "argocd@cluster" -f argocd-deploy-key -N ""
# Add argocd-deploy-key.pub to GitHub repository deploy keys
argocd repo add [email protected]:USERNAME/gitops-repo.git \
--ssh-private-key-path argocd-deploy-key \
--name gitops-repo
# Verify repository connection
argocd repo list
# Configure Ingress for UI (optional)
cat <<EOF | kubectl apply -f -
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
name: argocd-server-ingress
namespace: argocd
annotations:
cert-manager.io/cluster-issuer: letsencrypt-prod
nginx.ingress.kubernetes.io/ssl-passthrough: "true"
nginx.ingress.kubernetes.io/backend-protocol: "HTTPS"
spec:
ingressClassName: nginx
tls:
- hosts:
- argocd.example.com
secretName: argocd-tls
rules:
- host: argocd.example.com
http:
paths:
- path: /
pathType: Prefix
backend:
service:
name: argocd-server
port:
number: 443
EOF
→ Argo CD in argocd NS. UI via port-forward or Ingress. Admin pw changed. Repo added (SSH/token). Connection verified.
If err: CrashLoopBackOff → kubectl logs -n argocd -l app.kubernetes.io/name=argocd-server. Repo fail → verify token has repo access or SSH key in deploy keys. Ingress SSL → ensure cert-manager issued. Login fail → retrieve pw again or kubectl delete secret argocd-initial-admin-secret -n argocd + restart.
Step 2: App manifest + deploy first
# Create Git repository structure
mkdir -p gitops-repo/{apps,infra,projects}
cd gitops-repo
# Create sample application
mkdir -p apps/myapp/overlays/{dev,staging,prod}
mkdir -p apps/myapp/base
# Base Kustomization
cat > apps/myapp/base/kustomization.yaml <<EOF
apiVersion: kustomize.config.k8s.io/v1beta1
kind: Kustomization
resources:
- deployment.yaml
- service.yaml
EOF
cat > apps/myapp/base/deployment.yaml <<EOF
apiVersion: apps/v1
kind: Deployment
metadata:
name: myapp
spec:
replicas: 3
selector:
matchLabels:
app: myapp
template:
metadata:
labels:
app: myapp
spec:
containers:
- name: myapp
image: ghcr.io/username/myapp:v1.0.0
ports:
- containerPort: 8080
EOF
cat > apps/myapp/base/service.yaml <<EOF
apiVersion: v1
kind: Service
metadata:
name: myapp
spec:
selector:
app: myapp
ports:
- port: 80
targetPort: 8080
EOF
# Production overlay
cat > apps/myapp/overlays/prod/kustomization.yaml <<EOF
apiVersion: kustomize.config.k8s.io/v1beta1
kind: Kustomization
namespace: production
resources:
- ../../base
replicas:
- name: myapp
count: 5
images:
- name: ghcr.io/username/myapp
newTag: v1.0.0
EOF
# Commit to Git
git add .
git commit -m "Add myapp application manifests"
git push
# Create Argo CD Application
cat > argocd-apps/myapp-prod.yaml <<EOF
apiVersion: argoproj.io/v1alpha1
kind: Application
metadata:
name: myapp-prod
namespace: argocd
finalizers:
- resources-finalizer.argocd.argoproj.io
spec:
project: default
source:
repoURL: https://github.com/USERNAME/gitops-repo
targetRevision: main
path: apps/myapp/overlays/prod
destination:
server: https://kubernetes.default.svc
namespace: production
syncPolicy:
automated:
prune: true # Delete resources removed from Git
selfHeal: true # Auto-sync on drift detection
allowEmpty: false
syncOptions:
- CreateNamespace=true
- PruneLast=true
retry:
limit: 5
backoff:
duration: 5s
factor: 2
maxDuration: 3m
revisionHistoryLimit: 10
EOF
# Apply Application via kubectl
kubectl apply -f argocd-apps/myapp-prod.yaml
# Or create via CLI
argocd app create myapp-prod \
--repo https://github.com/USERNAME/gitops-repo \
--path apps/myapp/overlays/prod \
--dest-server https://kubernetes.default.svc \
--dest-namespace production \
--sync-policy automated \
--auto-prune \
--self-heal
# Watch sync status
argocd app get myapp-prod --watch
# Verify application
kubectl get all -n production
argocd app sync myapp-prod # Manual sync if automated disabled
→ App auto-synced from Git. Resources in prod NS. UI healthy. Auto sync w/ prune + self-heal. Sync succeeds within retry.
If err: sync fail → argocd app get myapp-prod + kubectl get events -n production. Kustomize build err → test local kustomize build apps/myapp/overlays/prod. NS err → verify exists or enable CreateNamespace. Pruning issues → check finalizers + owner refs.
Step 3: App-of-apps for multi-env
# Create app-of-apps structure
mkdir -p argocd-apps/{projects,infra,apps}
# Define projects for RBAC
cat > argocd-apps/projects/production.yaml <<EOF
apiVersion: argoproj.io/v1alpha1
# ... (see EXAMPLES.md for complete configuration)
→ Root app manages all children. New apps auto-deployed on Git add. Infra apps before app apps (sync waves). Projects enforce RBAC. Tree shows parent-child.
If err: circular deps → sync waves control order. Project perm errs → verify sourceRepos + destinations match. Recursive dir issues → valid YAML no conflict. Missing children → argocd app get root-app.
Step 4: Image Updater for auto deploy
# Install Argo CD Image Updater
kubectl apply -n argocd -f https://raw.githubusercontent.com/argoproj-labs/argocd-image-updater/stable/manifests/install.yaml
# Configure image update strategy via annotations
cat > argocd-apps/myapp-prod-autoupdate.yaml <<EOF
apiVersion: argoproj.io/v1alpha1
# ... (see EXAMPLES.md for complete configuration)
→ Updater monitors registry for new images matching tag patterns. Semver → latest stable. Git commits auto w/ new tags. Apps sync w/ updates. Staging uses digest (immutable).
If err: registry access → verify image-updater has pull creds (secret or SA). Write-back fail → check git-creds push perms. No updates → verify regex matches w/ argocd-image-updater test ghcr.io/username/myapp. Auth → check image-updater logs.
Step 5: Progressive delivery via Argo Rollouts
# Install Argo Rollouts controller
kubectl create namespace argo-rollouts
kubectl apply -n argo-rollouts -f https://github.com/argoproj/argo-rollouts/releases/latest/download/install.yaml
# Install Rollouts kubectl plugin
curl -LO https://github.com/argoproj/argo-rollouts/releases/latest/download/kubectl-argo-rollouts-linux-amd64
# ... (see EXAMPLES.md for complete configuration)
→ Progressive traffic shift to canary. Analysis validates success rate each step. Auto-promote on success, rollback on fail. Argo CD syncs Rollouts. Dashboard shows real-time progress.
If err: analysis fails → verify Prometheus accessible + query valid. Traffic routing → Ingress annotations + canary service endpoints. Stuck → manual promote/abort. Revision mismatch → ensure Argo CD sync policy no conflict w/ Rollouts controller.
Step 6: Drift detection + webhook notifications
# Configure drift detection in Application
cat > argocd-apps/myapp-strict.yaml <<EOF
apiVersion: argoproj.io/v1alpha1
kind: Application
metadata:
name: myapp-prod
# ... (see EXAMPLES.md for complete configuration)
→ Self-heal auto-reverts manual kubectl changes. Slack notif on sync fail + successful deploys. Webhooks trigger external (PagerDuty, monitoring, ITSM). Drift alerts show changes + who (via Git history).
If err: self-heal not triggering → verify auto sync enabled + refresh interval not too long (default 3m). Notif fails → test Slack token w/ curl + bot added to channels. Ignored diffs not working → check JSON pointer syntax. Webhook errs → endpoint accessibility + auth headers.
Check
- Argo CD / Flux installed + accessible
- Git repo connected w/ auth
- Apps auto-sync from Git on commit
- Manual kubectl changes reverted by self-heal
- App-of-apps deploys multiple
- Image Updater promotes by tag patterns
- Argo Rollouts progressive canary
- Notifications to Slack/email on sync
- Drift detection alerts out-of-band
- RBAC enforces project access
Traps
- Auto prune disabled: removed from Git remain in cluster.
prune: truein sync. - No sync waves: infra deployed after dependent apps.
argocd.argoproj.io/sync-waveannotations. - Ignoring HPA replicas: sync fails b/c HPA changed count. Add
/spec/replicasto ignoreDifferences. - Write-back conflicts: Updater commits conflict w/ manual. Separate branch or fine RBAC.
- Missing finalizers: deletion leaves orphans. Add
resources-finalizer.argocd.argoproj.io. - No analysis templates: Rollouts promote w/o validation. Impl AnalysisTemplates w/ metrics queries.
- Secrets in Git: plaintext committed. Use Sealed Secrets / External Secrets.
- Self-heal too aggressive: reverts legitimate emergency changes. Annotations temporarily disable or approval gates.
→
configure-git-repository— repo structure for GitOpsmanage-git-branches— branch strategies for env promotiondeploy-to-kubernetes— K8s resources managed by GitOpsmanage-kubernetes-secrets— Sealed Secrets integrationbuild-ci-cd-pipeline— CI builds, GitOps deployssetup-container-registry— image promotion between registries
GitHub 저장소
연관 스킬
executing-plans
디자인executing-plans 스킬은 검토 체크포인트가 포함된 통제된 배치로 실행할 완전한 구현 계획이 있을 때 사용합니다. 이 스킬은 계획을 불러와 비판적으로 검토한 후, 소규모 배치(기본값 3개 작업)로 작업을 실행하면서 각 배치 사이에 진행 상황을 아키텍트 검토를 위해 보고합니다. 이를 통해 내재된 품질 관리 체크포인트를 갖춘 체계적인 구현이 보장됩니다.
requesting-code-review
디자인이 스킬은 코드 변경 사항을 요구 사항에 따라 분석하기 위해 코드 리뷰어 하위 에이전트를 호출합니다. 작업 완료 후, 주요 기능 구현 후, 또는 메인 브랜치에 병합하기 전에 사용해야 합니다. 이 리뷰는 현재 구현체와 원래 계획을 비교하여 문제를 조기에 발견하는 데 도움이 됩니다.
connect-mcp-server
디자인이 스킬은 개발자들이 HTTP, stdio 또는 SSE 전송 방식을 통해 MCP 서버를 Claude Code에 연결하는 포괄적인 가이드를 제공합니다. GitHub, Notion 및 사용자 정의 API와 같은 외부 서비스를 통합하기 위한 설치, 구성, 인증 및 보안을 다룹니다. MCP 통합 설정, 외부 도구 구성 또는 Claude의 모델 컨텍스트 프로토콜 작업 시 활용하세요.
web-cli-teleport
디자인이 스킬은 작업 분석을 기반으로 개발자가 Claude Code 웹 인터페이스와 CLI 인터페이스 중 선택할 수 있도록 돕고, 두 환경 간 원활한 세션 텔레포트를 가능하게 합니다. 웹, CLI 또는 모바일 환경 전환 시 세션 상태와 컨텍스트를 관리하여 워크플로를 최적화합니다. 다양한 단계에서 서로 다른 도구가 필요한 복잡한 프로젝트에 사용하세요.
