The application itself is covered elsewhere, in Back End and Front End. This site is everything else you need before anyone can actually use what you built.
It reads as the path a project takes to production. Get it into git, get it onto a machine, put a server in front of it, secure the connection, then automate the whole thing so the next deploy is not manual. The AWS folder is the largest single group because that is where most of it ended up running.
Git and GitHub
| Git |
The tool itself |
| GitHub |
The longest page on the site, covering the platform and its automation |
| GitHub Workflows |
How Actions workflows are structured, how to reuse logic, and CI/CD pipelines for Django, Node, React, and Docker projects |
| GitHub Runners |
The machines that execute jobs. Hosted runners get a fresh VM per job with maintenance handled; self-hosted does not |
| The .github Folder |
What the special directory configures, and how to lay it out |
| PIP and Git Release |
Cutting a release to both PyPI and GitHub |
AWS
Sixteen pages, roughly one per service, each covering what it does and the trade-off or pitfall specific to it. AWS Overview is the map.
| AWS Overview |
How the services fit together, and where the platform is heading |
| Boto3 |
The Python SDK, and the longest page in this folder |
| AWS CLI |
AWS as a programmable terminal, which is where most automation starts |
| EC2 |
Instances, best practice, and how you would actually run Django on one |
| ECS |
Container orchestration, the managed way |
| EKS |
Kubernetes with Django on top: architecture, trade-offs, and the operational concerns |
| S3 |
Object storage |
| RDS |
Managed relational databases |
| VPC |
Your own isolated network: IP ranges, subnets, routing, and security groups |
| Route 53 |
DNS, record management, health checks, and domain registration |
| CloudFront |
The CDN, its edge locations, and its integration with S3, ALB, API Gateway, and Lambda@Edge |
| IAM |
Identity and access management, which is the part that gates everything else |
| Secrets Manager |
Storing credentials somewhere other than the repo |
| SES |
Sending transactional and bulk email, with tracking, templates, and dedicated IPs |
| SNS |
Pub/sub notifications |
| Bedrock |
Generative AI through one API across many model providers, with guardrails, knowledge bases for RAG, agents, fine-tuning, and VPC endpoints |
| Oracle Cloud |
Brief notes on the alternative |
Serving and Securing
| Nginx |
The web server and reverse proxy that sits in front of everything |
| Gunicorn |
The WSGI server that actually runs a Django or DRF application behind it |
| Certbot |
Automating Let’s Encrypt certificate issuance and renewal, which is what makes HTTPS free |
Networking and Remote Access
| iptables |
Linux packet filtering, at length |
| SSH, VNC and FTP |
The three ways into a remote machine |
| WireGuard |
A modern VPN on an EC2 instance, used here to reach a home network from outside |
| OpenVPN |
The older alternative, same goal |
| Wireshark |
Packet capture, for when nothing else explains it |
Infrastructure as Code
| Terraform |
Declarative infrastructure in HCL |
| Ansible |
Configuration management, the other half of the pair |
| Pulumi |
Terraform’s model expressed in a real programming language. The honest verdict is that its two genuine differentiators are the Automation API and testability, and the rest is parity with a smaller ecosystem |
Automation and Operations
| n8n |
Self-hostable workflow automation, a developer-facing alternative to Zapier with first-class API support |
| Prefect |
Orchestrating data workflows in Python |
| Event-Condition-Action |
The paradigm underneath rule engines and automation systems: on an event, check a condition, then act |
| Business Rule Engines |
Pulling business logic out of application code and treating rules as data rather than code |
| LGTM Stack |
Observability with OpenTelemetry plus Loki, Grafana, Tempo, and Mimir |
| Makefile |
A stub; the real treatment is in Software Tools |
Running a Product
| Security Compliance |
The clearest-eyed page here. Compliance is a sales function, not a security function. A certificate proves to procurement that controls exist and are audited; it does not make you secure, and confusing the two is how a company spends heavily on SOC 2 and still ships a tenant-isolation bug |
| Stripe |
Payment handling for a SaaS |
| Weather API |
Open-Meteo, which needs no API key, as a worked example of consuming a third-party service |
Not Covered Yet
- No Kubernetes page of its own. EKS covers the managed AWS route, but nothing explains Kubernetes itself. Container work generally lives in IaaS.
- Azure and GCP are absent, so the cloud coverage is AWS plus a brief Oracle note.
- No load testing or capacity planning, despite the serving pages assuming production traffic.
- Nothing on incident response: observability is covered, but not what to do when it fires.
- Makefile and Wireshark are stubs, at three and four cells.
- The practice-management API integrations are kept unlisted, so the worked client examples behind the blog post on healthcare APIs are not visible here.
Back to top