Last deploy: passed · 5/5 checks

Infrastructure that ships itself.

CloudPilot is an end-to-end DevOps pipeline: Terraform provisions the servers, Ansible configures them, GitHub Actions tests and deploys on every push, and Prometheus watches to make sure the deploy actually worked.

Explore the repo See how it deploys

PIPELINE.SCHEMATIC — main branch REV 1.0

Bill of materials

What's actually running

Five tools, each doing exactly one job — no orchestration platform hiding what's happening underneath.

Layer

Component

Role

Terraform

Provisions the VPC, subnets, and Auto Scaling Group on AWS

Docker

Packages the app into a small, non-root, health-checked image

GitHub Actions

Runs tests, validates infra, builds the image, triggers deploy

Ansible

Rolls the new image out to every instance and checks it's healthy

Prometheus / Grafana

Watches request volume and uptime after the deploy ships

On every push to main

Five stages, one command: git push

Each stage gates the next — a failed test or a bad Terraform plan stops the pipeline before anything touches production.

STAGE 01

Test

pytest runs against the app. Fails here, nothing downstream runs.

STAGE 02

Validate

terraform fmt and validate catch broken or malformed infra code.

STAGE 03

Build

Docker image is built and pushed to GitHub Container Registry.

STAGE 04

Deploy

Ansible rolls the new image out across the EC2 fleet.

STAGE 05

Verify

Health checks confirm the new version is actually serving traffic.

Under the hood

A VPC across two availability zones, not a single point of failure

Public subnets in two AWS availability zones hold an Auto Scaling Group running the containerized app behind a security group that only opens port 80 to the world and port 22 to one admin IP.

Read the full architecture doc →