Orchestrating Deployment Pipelines
Overview
Orchestrate multi-stage deployment pipelines that coordinate builds, tests, approvals, and releases across environments (dev, staging, production). Implement deployment strategies including blue-green, canary, rolling updates, and feature flags using Kubernetes, cloud-native services, and CI/CD platforms.
Prerequisites
- CI/CD platform configured (GitHub Actions, GitLab CI, Jenkins, ArgoCD)
- Kubernetes cluster with
kubectl access or cloud deployment target (ECS, Cloud Run, App Engine)
- Container registry with built and tagged images ready for deployment
- Environment-specific configuration (secrets, environment variables) stored securely
- Monitoring and alerting configured to detect deployment failures
Instructions
- Define the deployment topology: target environments, promotion flow (dev -> staging -> production), and approval gates
- Select deployment strategy per environment: rolling update for staging, canary or blue-green for production
- Generate deployment manifests (Kubernetes Deployments, Services, Ingress) or cloud service configurations
- Implement pre-deployment checks: database migration status, dependency health, configuration validation
- Configure canary analysis: route 5-10% of traffic to new version, monitor error rate and latency for 15 minutes before full rollout
- Add post-deployment verification: smoke tests, health check endpoints, synthetic monitoring
- Implement automated rollback triggers: revert if error rate exceeds 1% or P99 latency doubles during canary phase
- Set up deployment notifications: Slack messages with deployment status, version, environment, and commit link
- Document the deployment runbook with manual intervention procedures for edge cases
Output
- Deployment pipeline configurations (GitHub Actions workflows, ArgoCD Applications)
- Kubernetes manifests with deployment strategy annotations
- Canary analysis configuration (Flagger, Argo Rollouts)
- Pre/post-deployment hook scripts
- Deployment runbook with rollback procedures
Error Handling
| Error |
Cause |
Solution |
ImagePullBackOff |
Image tag not found in registry or auth failure |
Verify image exists with docker manifest inspect; check imagePullSecrets |
CrashLoopBackOff |
Application failing to start in new version |
Check pod logs with kubectl logs; verify environment variables and config maps |
Canary analysis failed |
Error rate or latency exceeded threshold during canary |
Automatic rollback triggered; investigate logs from canary pods before retrying |
Deployment stuck in Progressing |
Insufficient resources or pod scheduling failure |
Check kubectl describe deployment for events; verify resource requests and node capacity |
Database migration failed |
Schema conflict or lock timeout |
Run migrations independently before deployment; add retry logic and connection timeout |
Examples
- "Create a deployment pipeline that builds on PR merge, deploys to staging automatically, runs integration tests, then requires manual approval for production with canary rollout."
- "Set up Argo Rollouts for a Kubernetes deployment with 10% canary traffic, Prometheus-based analysis, and automatic rollback on error rate > 0.5%."
- "Generate a blue-green deployment for an ECS service with ALB target group switching and automatic rollback on health check failure."
Resources
Source: jeremylongshore/claude-code-plugins-plus-skills → skills/.curated/orchestrating-deployment-pipelines/SKILL.md
Also appears in: jeremylongshore/claude-code-plugins-plus-skills/plugins/devops/deployment-pipeline-orchestrator/skills/orchestrating-deployment-pipelines/SKILL.md
1---2name: orchestrating-deployment-pipelines3description: 'Deploy use when you need to work with deployment and CI/CD. This skill provides deployment automation and orchestration with comprehensive guidance and automation. Trigger with phrases like "deploy application", "create pipeline", or "automate deployment". '4---5
6# Orchestrating Deployment Pipelines
7
8## Overview
9
10Orchestrate multi-stage deployment pipelines that coordinate builds, tests, approvals, and releases across environments (dev, staging, production). Implement deployment strategies including blue-green, canary, rolling updates, and feature flags using Kubernetes, cloud-native services, and CI/CD platforms.
11
12## Prerequisites
13
14- CI/CD platform configured (GitHub Actions, GitLab CI, Jenkins, ArgoCD)
15- Kubernetes cluster with `kubectl` access or cloud deployment target (ECS, Cloud Run, App Engine)
16- Container registry with built and tagged images ready for deployment
17- Environment-specific configuration (secrets, environment variables) stored securely
18- Monitoring and alerting configured to detect deployment failures
19
20## Instructions
21
221. Define the deployment topology: target environments, promotion flow (dev -> staging -> production), and approval gates
232. Select deployment strategy per environment: rolling update for staging, canary or blue-green for production
243. Generate deployment manifests (Kubernetes Deployments, Services, Ingress) or cloud service configurations
254. Implement pre-deployment checks: database migration status, dependency health, configuration validation
265. Configure canary analysis: route 5-10% of traffic to new version, monitor error rate and latency for 15 minutes before full rollout
276. Add post-deployment verification: smoke tests, health check endpoints, synthetic monitoring
287. Implement automated rollback triggers: revert if error rate exceeds 1% or P99 latency doubles during canary phase
298. Set up deployment notifications: Slack messages with deployment status, version, environment, and commit link
309. Document the deployment runbook with manual intervention procedures for edge cases
31
32## Output
33
34- Deployment pipeline configurations (GitHub Actions workflows, ArgoCD Applications)
35- Kubernetes manifests with deployment strategy annotations
36- Canary analysis configuration (Flagger, Argo Rollouts)
37- Pre/post-deployment hook scripts
38- Deployment runbook with rollback procedures
39
40## Error Handling
41
42| Error | Cause | Solution |
43|-------|-------|---------|
44| `ImagePullBackOff` | Image tag not found in registry or auth failure | Verify image exists with `docker manifest inspect`; check `imagePullSecrets` |
45| `CrashLoopBackOff` | Application failing to start in new version | Check pod logs with `kubectl logs`; verify environment variables and config maps |
46| `Canary analysis failed` | Error rate or latency exceeded threshold during canary | Automatic rollback triggered; investigate logs from canary pods before retrying |
47| `Deployment stuck in Progressing` | Insufficient resources or pod scheduling failure | Check `kubectl describe deployment` for events; verify resource requests and node capacity |
48| `Database migration failed` | Schema conflict or lock timeout | Run migrations independently before deployment; add retry logic and connection timeout |
49
50## Examples
51
52- "Create a deployment pipeline that builds on PR merge, deploys to staging automatically, runs integration tests, then requires manual approval for production with canary rollout."
53- "Set up Argo Rollouts for a Kubernetes deployment with 10% canary traffic, Prometheus-based analysis, and automatic rollback on error rate > 0.5%."
54- "Generate a blue-green deployment for an ECS service with ALB target group switching and automatic rollback on health check failure."
55
56## Resources
57
58- Kubernetes deployment strategies: https://kubernetes.io/docs/concepts/workloads/controllers/deployment/
59- Argo Rollouts: https://argoproj.github.io/argo-rollouts/
60- Flagger (progressive delivery): https://flagger.app/
61- AWS ECS blue-green: https://docs.aws.amazon.com/AmazonECS/latest/developerguide/deployment-type-bluegreen.html
62
63---
64
65**Source:** [`jeremylongshore/claude-code-plugins-plus-skills`](https://github.com/jeremylongshore/claude-code-plugins-plus-skills) → `skills/.curated/orchestrating-deployment-pipelines/SKILL.md`
66
67**Also appears in:** `jeremylongshore/claude-code-plugins-plus-skills/plugins/devops/deployment-pipeline-orchestrator/skills/orchestrating-deployment-pipelines/SKILL.md`
thedixitjain