sutanunandigrami/infra-deploy

Infrastructure Workflows

Terraform

1. terraform fmt -recursive — format all .tf files
2. tflint — lint for errors and best practices
3. terraform validate — syntax validation
4. terraform plan -out=plan.tfplan — preview changes (ALWAYS do this first)
5. infracost breakdown --path=plan.tfplan — estimate cost impact
6. Only after review: terraform apply plan.tfplan

NEVER run terraform apply -auto-approve or terraform destroy without explicit operator approval.

Ansible

1. ansible-lint — lint playbooks
2. ansible-playbook --check -i inventory site.yml — dry run
3. ansible-playbook -i inventory site.yml — actual run

Docker

1. hadolint Dockerfile — lint before building
2. docker build -t <name>:<tag> . — build image
3. trivy image <name>:<tag> — scan for vulnerabilities
4. dive <name>:<tag> — analyze layer efficiency
5. docker compose up -d — deploy

Kubernetes

1. kubectl get pods -A — cluster overview
2. k9s — interactive management
3. stern <pod-prefix> — tail logs from multiple pods
4. helm list -A — check installed charts

Hetzner Cloud

Run hcloud --help for available commands. Common: hcloud server list, hcloud server create, hcloud firewall list

Rules

• ALWAYS plan before apply
• ALWAYS scan images before deploying
• ALWAYS lint IaC files before committing
• Use sops or age for secrets, never plaintext