DevOps Team Policy

The eight DevOps phases, tools per phase, and responsibilities.

This is a generalized view of the eight DevOps phases in plain language for anyone taking up a DevOps role on a project. The DevOps team is involved from project inception through maintenance, complementing the dev, testing, operations, security, and database teams to reduce the gap between development and operations and enable faster release cycles.

Deployment pipeline

The eight phases

#	Phase	What we do	Tools
1	Plan	Requirements, architecture & design, timelines, account creation	GitHub, AWS, Figma, MongoDB Atlas, OpenAI, etc.
2	Code	Development, check-in, static analysis, API testing, CI checks	GitHub, VS Code, SonarQube, Axiom, Sentry
3	Build & Push / CI	Build code, build & push Docker image, image scanning	Node, Docker, ECR
4	Continuous Deployment	Task definitions, deploy to dev → staging, TLS certs, hosted zones	GitHub Workflows, ECS
5	Reliability & Scaling	ALB, target groups, ECS autoscaling, uptime monitoring, branch rules	Auto Scaling Group, Load Balancer, Pingdom
6	Security	Secrets Manager, child keys, private networking, encryption, IP allow-lists	AWS security services
7	Testing	Load & stress, functional, UI, UAT, runtime error finding	Locust, Sentry
8	Release / Operations	Release notes, version tagging, code freeze, go-live, rollback, hotfix	GitHub Release Tags
9	Monitoring	Monitor infrastructure and application resources	CloudWatch

Phase detail

1. Plan

Create required accounts (GitHub, AWS, Figma, MongoDB Atlas, Vercel, Apple/Play developer, OpenAI, GoDaddy, LangChain, Pingdom, etc.), provision infrastructure (ECS/EKS, SNS, SES, autoscaling, databases, repositories, Cognito, IAM, SonarQube, Prometheus, Grafana), install required tooling (kubectl, eksctl, AWS CLI, Node/NPM, Docker, pgAdmin, Mongo Compass, Postman), and enable MFA for all accounts.

2. Code

Develop locally, test before pushing, and commit to GitHub. SonarQube performs static analysis on push to the alpha branch; code must pass the quality gates before progressing.

3. Build & Push

Raise a PR against the alpha branch. SonarQube runs automatically; once it passes, merge. Adding the Dockerfile and GitHub workflow files triggers a pipeline that builds the image and pushes it to the registry. See Docker for image details.

4. Deployment (alpha, beta, prod)

Ensure images are free of vulnerabilities, create Helm charts (deployment, svc, ingress, sa), set up ArgoCD for resource status and logs, create a Route 53 hosted zone, issue ACM TLS certificates, and map DNS. Roll back to the previous version if issues occur, then fix and redeploy.

5. Production best practices

VPC peering between MongoDB Atlas and AWS, autoscaling, alerts, OpenTelemetry tracing, delete protection, CloudWatch logs and alarms, continuous backups, and Pingdom endpoint monitoring. See Production Best Practices.

6. Security best practices

Block developer access to beta/production branches, branch protection, HTTP→HTTPS redirect, avoid root keys (use limited child keys), private networking, disable public access, encrypt data at rest (KMS) and in transit, encrypt snapshots, rotate admin passwords, and run GitHub checks before merge. See the Security section.