Intermediate / Advanced · ~60–70 hours
AI-Enabled DevOps SDLC from Cursor
Run the entire software delivery lifecycle for a production-grade microservices app without leaving your IDE — from configuring Cursor rules and AI agents, through CI/CD and ArgoCD deployments to Kubernetes, to troubleshooting production bugs via CloudWatch and shipping hotfixes.
Who this is for
Backend, platform, and DevOps engineers who want one IDE for the whole SDLC.
You already ship prod work. This course shows you how to consolidate your tooling into a single Cursor workflow powered by AI agents — not how to learn Kubernetes or Docker from scratch.
Prerequisites
- ✓ Working knowledge of at least one backend language in the stack (Node.js, Java, Python, or C#)
- ✓ Comfortable with Git workflows — branching, PRs, rebasing
- ✓ Basic Docker understanding — images, containers, volumes
- ✓ AWS account with programmatic access and Kubernetes fundamentals (pods, deployments, services)
- ✓ Cursor IDE installed with an active subscription — AI features are used throughout
What you'll build
The Pretzel Shop — polyglot microservices, full SDLC pipeline.
A microservices e-commerce app with four backend services in four languages, plus PostgreSQL, Redis, and a React + Vite frontend. Around it, you build the entire delivery pipeline from inside Cursor.
Four-language microservices
Node.js / Express, Java / Spring Boot, Python / FastAPI, C# / .NET 9 — each with its own service boundary and conventions.
Cursor rules, skills & agents
Language-specific rules, scaffolding skills, architecture reviewer, best-practices auditor, test writer, and release certifier subagents.
MCP integrations
Jira, Confluence, AWS, and Terraform connected via MCP — query boards, file tickets, and look up docs without leaving the IDE.
ArgoCD & k6
GitHub Actions → ECR → EKS via ArgoCD. k6 load tests prove zero downtime during rolling updates and gate promotion.
- →Pre-commit hooks and Cursor hooks — lint, scan secrets, validate Dockerfiles, enforce conventional commits
- →Twelve-gate release certification pipeline — blocks production until quality, security, and ops checks pass
- →CloudWatch Logs Insights and OpenTelemetry trace correlation for production incident diagnosis
- →Bugbot pre-merge review and Automations to triage failed CI — all triggered from Cursor
Roadmap
16 sections across three parts.
Part I · Foundation
Configuring Cursor for AI-Enabled DevOps
Workspace settings, Agent modes, `.cursor/` folder structure, project-wide and language-specific rules.
Running Pretzel Shop Microservices Locally
Docker Compose for all four backends, PostgreSQL, Redis, and frontend — verify health endpoints and inter-service communication.
Writing Rules and Skills for Expert Codebase Updates
`AGENTS.md`, language-specific rules, and reusable skills for endpoints, tests, and Dockerfiles.
Connecting Cursor to Jira, Confluence, AWS, Terraform, and CLI Tools
Atlassian MCP, AWS MCP proxy, Terraform MCP, plus `gh`, `aws`, `terraform`, and `kubectl` from the IDE.
Verifying Connections with Lite Tests
`verify-connections.sh` — pass/fail report for every integration, wired into workspace open tasks.
Part II · Enabling SDLC
Enforcing Pre-commit Hooks and Custom Cursor Hooks
pre-commit framework across four languages, MCP audit gates, secret redaction, and Automations.
Writing .cursor/agents for Architecture, Best Practices, and Testing
Architecture reviewer, best-practices auditor, and test writer subagents with fixture-based validation.
Custom Skills for Planning, Coding, Dockerfiles, Pipelines, and Terraform
Six skills including `full-feature` — chains planning through infrastructure in one orchestrated flow.
Testing Skills and Deploying via ArgoCD from Cursor
End-to-end deploy: AI-authored code → Bugbot → GitHub Actions → ECR → ArgoCD sync on staging.
Zero-Downtime Validation with Load Testing
k6 suite during ArgoCD rolling updates — error rate < 0.1%, p95 < 500ms, tuned `maxSurge`/`maxUnavailable`.
Best Practices and ADRs
Seven MADR 4.0 ADRs — OpenTelemetry, probes, multi-stage Dockerfiles, rollback under load, additive migrations.
Release Certification Agents
Twelve-gate certification pipeline — linters, Trivy, Grype, rollback under load, zero-downtime proof, Bugbot.
Part III · Mastery
Troubleshooting a Production Bug from Cursor
Debug Mode, CloudWatch Logs Insights, OpenTelemetry traces, Jira ticket via Atlassian MCP — all from the IDE.
Implementing the Fix from Cursor
Plan Mode, git worktree, coding skill, regression tests, Bugbot `/review`, PR linked to Jira, certification report.
Deploying the Fix to the Lower Environment
`gh run watch`, ArgoCD sync, k6 during rollout, CloudWatch verification, Jira updated to "Staging Verified."
Preparing the Release and Deploying to Production
semantic-release, production ArgoCD sync, k6 smoke test, Jira closed with deployment evidence, Confluence incident page.
Tech stack
Everything you touch in the course.
IDE
Cursor, MCP, Bugbot
Languages
Node, Java, Python, C#
Frontend
React, Vite
Containers
Docker, Compose
Orchestration
Kubernetes (EKS)
GitOps
ArgoCD
CI/CD
GitHub Actions
IaC
Terraform
Data
RDS Postgres, Redis
Observability
OTel, CloudWatch
Load testing
k6
Security
Trivy, Grype, hadolint
PM
Jira, Confluence
Release
semantic-release
CLI
aws, gh, kubectl, tf
Learning outcomes
What you'll be able to do.
- ✓Configure Cursor as a single-pane IDE — rules, skills, agents, hooks, MCP servers, and terminal profiles for a polyglot project
- ✓Build reusable skills that scaffold endpoints, generate Dockerfiles, update GitHub Actions workflows, and create Terraform modules
- ✓Create specialized subagents for architecture review, best-practices auditing, test generation, and release certification
- ✓Connect Cursor to Jira, Confluence, AWS, and Terraform via MCP — manage tickets and docs without leaving the IDE
- ✓Deploy microservices to EKS through GitHub Actions and ArgoCD with automated and manual sync policies
- ✓Write k6 load tests that prove zero downtime during rolling updates and gate promotion on threshold results
- ✓Build a twelve-gate release certification pipeline that blocks production until quality, security, and ops checks pass
- ✓Run the full incident lifecycle — CloudWatch detection, Jira ticket, fix, staging validation, production deploy, ticket closure — from one IDE session
Ready to run your SDLC from Cursor?
~20 minutes on a call. I'll walk you through the course structure and whether it fits where you are. No pitch deck — just a straight conversation.
Book a call →Best fit if you already ship prod work and want to consolidate your SDLC into a single AI-powered IDE workflow.
The differentiator
The IDE is the cockpit.
Most courses teach tools in isolation — a Kubernetes module here, a CI/CD module there. This course teaches one workflow: everything runs from inside Cursor.
The full loop you'll operate end-to-end:
No browser tab for AWS console. No separate terminal for `kubectl`. No context-switching to Jira. The IDE is your single control plane for the entire software delivery lifecycle.