Patterns d'Orchestration DevOps

VérifiéSûr

Ce skill fournit des modèles d'orchestration DevOps pour configurer des pipelines CI/CD, effectuer des déploiements sans interruption, gérer l'infrastructure en tant que code (Terraform, Kubernetes), et mettre en place la surveillance et la réponse aux incidents. Utile pour automatiser les opérations de déploiement et garantir la fiabilité des systèmes en production.

Spar Skills Guide Bot
DevOpsIntermédiaire
14002/06/2026
Claude Code
#ci-cd#deployment#infrastructure-as-code#monitoring#incident-response

Recommandé pour

Notre avis

Cette compétence orchestre des workflows DevOps complexes en décomposant les tâches en phases exploratoires, planifiées et parallélisées, couvrant CI/CD, déploiement sans interruption, infrastructure, surveillance et réponse aux incidents.

Points forts

  • Automatisation complète du cycle DevOps avec des patrons reproductibles
  • Parallélisation efficace des tâches pour réduire les temps d'exécution
  • Gestion native des déploiements sans temps d'arrêt et des migrations de bases de données
  • Intégration de la sécurité et de la surveillance dans le pipeline

Limites

  • Nécessite une infrastructure cloud ou Kubernetes préexistante pour certains patrons
  • Peut nécessiter des ajustements manuels pour des cas spécifiques non couverts
  • Dépend de la capacité à modéliser correctement l'infrastructure existante
Quand l'utiliser

Utilisez cette compétence pour industrialiser des déploiements complexes, automatiser des pipelines CI/CD ou mettre en place une infrastructure as code robuste.

Quand l'éviter

Évitez-la pour des projets très simples ou monolithiques où la surcharge d'orchestration n'est pas justifiée.

Analyse de sécurité

Sûr
Score qualité90/100

The skill provides only patterns and guidance; no destructive or exfiltrating commands are present. The mention of npx cc-mirror is only an illustrative example, not an instruction to run untrusted code.

Aucun point d'attention détecté

Exemples

Set up CI/CD pipeline
Set up CI/CD for this project
Zero-downtime deployment
Deploy the new version without downtime
Terraform AWS infrastructure
Create Terraform for our AWS setup

DevOps Orchestration Patterns

┌─────────────────────────────────────────────────────────────┐
│                                                             │
│   Infrastructure as carefully orchestrated as code.         │
│   Safe deployments. Fast pipelines. Reliable operations.    │
│                                                             │
└─────────────────────────────────────────────────────────────┘

Load when: CI/CD pipeline, deployment, infrastructure as code, monitoring, incident response Common patterns: Pipeline Setup, Zero-Downtime Deployment, Incident Triage

Table of Contents

  1. CI/CD Pipeline
  2. Deployment
  3. Infrastructure
  4. Monitoring and Alerting
  5. Incident Response

CI/CD Pipeline

Pattern: Pipeline Setup

User Request: "Set up CI/CD for this project"

Phase 1: EXPLORE
└─ Explore agent: Analyze project structure, build system, test setup

Phase 2: PLAN
└─ Plan agent: Design pipeline stages and requirements

Phase 3: FAN-OUT (Parallel stage implementation)
├─ Agent A: Build stage (compile, dependencies)
├─ Agent B: Test stage (unit, integration)
├─ Agent C: Security scan stage (SAST, dependencies)
└─ Agent D: Deploy stage (environments, rollback)

Phase 4: PIPELINE
└─ General-purpose agent: Wire stages, add notifications

Pattern: Pipeline Optimization

User Request: "Speed up our CI pipeline"

Phase 1: FAN-OUT (Parallel analysis)
├─ Explore agent: Analyze current pipeline duration by stage
├─ Explore agent: Find parallelization opportunities
├─ Explore agent: Check caching effectiveness
└─ Explore agent: Review resource allocation

Phase 2: REDUCE
└─ Plan agent: Prioritized optimization plan

Phase 3: FAN-OUT (Implement optimizations)
├─ Agent A: Implement caching improvements
├─ Agent B: Parallelize independent jobs
└─ Agent C: Optimize resource-heavy stages

Deployment

Pattern: Zero-Downtime Deployment

User Request: "Deploy the new version without downtime"

Phase 1: EXPLORE
└─ Explore agent: Current deployment strategy, infrastructure

Phase 2: PIPELINE
├─ General-purpose agent: Build and tag new image
├─ General-purpose agent: Deploy to canary/blue environment
├─ Background agent: Run smoke tests
└─ General-purpose agent: Shift traffic, drain old instances

Phase 3: BACKGROUND
└─ Background agent: Monitor error rates post-deploy

Pattern: Database Migration Deployment

Phase 1: PIPELINE (Safe migration)
├─ General-purpose agent: Backup current database
├─ General-purpose agent: Run forward migration
├─ General-purpose agent: Deploy compatible app version
└─ Background agent: Verify data integrity

Infrastructure

Pattern: Infrastructure as Code

User Request: "Create Terraform for our AWS setup"

Phase 1: EXPLORE
└─ Explore agent: Map current infrastructure, requirements

Phase 2: FAN-OUT (Parallel module creation)
├─ Agent A: Network module (VPC, subnets, security groups)
├─ Agent B: Compute module (ECS/EKS, EC2)
├─ Agent C: Database module (RDS, ElastiCache)
├─ Agent D: Storage module (S3, EFS)
└─ Agent E: CDN and DNS module

Phase 3: PIPELINE
├─ General-purpose agent: Wire modules, configure state backend
└─ Background agent: Validate and plan

Pattern: Kubernetes Manifest Generation

Phase 1: EXPLORE
└─ Explore agent: Analyze application requirements

Phase 2: FAN-OUT
├─ Agent A: Deployment manifests
├─ Agent B: Service and ingress manifests
├─ Agent C: ConfigMaps and Secrets
├─ Agent D: HPA and PDB configurations
└─ Agent E: NetworkPolicies

Monitoring and Alerting

Pattern: Observability Setup

User Request: "Set up monitoring for the application"

Phase 1: FAN-OUT (Parallel pillar implementation)
├─ Agent A: Metrics (Prometheus, CloudWatch)
├─ Agent B: Logging (ELK, CloudWatch Logs)
├─ Agent C: Tracing (Jaeger, X-Ray)
└─ Agent D: Dashboards (Grafana, custom)

Phase 2: PIPELINE
└─ General-purpose agent: Configure alerting rules, runbooks

Incident Response

Pattern: Incident Triage

User Request: "Production is down!"

Phase 1: FAN-OUT (Rapid parallel diagnosis)
├─ Agent A: Check application logs for errors
├─ Agent B: Check infrastructure metrics
├─ Agent C: Check recent deployments
├─ Agent D: Check external dependencies
└─ Agent E: Check database health

Phase 2: REDUCE (Fast)
└─ General-purpose agent: Identify most likely cause

Phase 3: PIPELINE
├─ General-purpose agent: Implement fix or rollback
└─ Background agent: Verify recovery

Pattern: Post-Incident Review

Phase 1: FAN-OUT (Evidence gathering)
├─ Explore agent: Timeline of events
├─ Explore agent: Relevant logs and metrics
├─ Explore agent: Changes before incident
└─ Explore agent: Response actions taken

Phase 2: PIPELINE
├─ General-purpose agent: Root cause analysis
├─ General-purpose agent: Impact assessment
└─ General-purpose agent: Action items and preventions

Phase 3: REDUCE
└─ General-purpose agent: Post-mortem document

Task Management for DevOps

Structure infrastructure work with safety checkpoints:

# Create DevOps tasks
npx cc-mirror tasks create --subject "Assess infrastructure" --description "Current state, requirements..."
npx cc-mirror tasks create --subject "Plan changes" --description "Design with minimal disruption..."
npx cc-mirror tasks create --subject "Implement network changes" --description "VPC, security groups..."
npx cc-mirror tasks create --subject "Implement compute changes" --description "ECS, scaling..."
npx cc-mirror tasks create --subject "Validate deployment" --description "Health checks, smoke tests..."
npx cc-mirror tasks create --subject "Configure monitoring" --description "Alerts, dashboards..."

# Sequential safety gates
npx cc-mirror tasks update 2 --add-blocked-by 1
npx cc-mirror tasks update 3 --add-blocked-by 2
npx cc-mirror tasks update 4 --add-blocked-by 2
npx cc-mirror tasks update 5 --add-blocked-by 3,4
npx cc-mirror tasks update 6 --add-blocked-by 5
# Parallel infrastructure implementation (sonnet for well-structured work)
Task(subagent_type="general-purpose", prompt="Task 3: Implement network changes...",
     model="sonnet", run_in_background=True)
Task(subagent_type="general-purpose", prompt="Task 4: Implement compute changes...",
     model="sonnet", run_in_background=True)

Safety Principles

  1. Always have rollback plan before deploying
  2. Background monitor during and after deployment
  3. Parallel diagnosis during incidents for speed
  4. Document everything for future incidents
  5. Test in staging before production

─── ◈ DevOps ────────────────────────────
Skills similaires