$ ls -t work/

Work

Newest first. Each ticket is a case study: what I built, what broke, what I would change.

TICKET 008CASE STUDY

Migrating an AEO Agency Off Squarespace

Rebuilt Penpixel Creative's own site from Squarespace into a custom Astro build on Cloudflare Pages, with a scoped CSP, HSTS and AI-crawler allowlisting.

Astro · Tailwind CSS · Cloudflare Pages · CSP · HSTS preload · Cloudflare TurnstileRead the case study →
TICKET 007CASE STUDY

28 AWS Resources for $22/Month

Auto-scaling, highly available AWS infrastructure in Terraform: ALB, two availability zones, CloudWatch scaling and S3 state with native locking.

Terraform · AWS EC2 · ALB · Auto Scaling · CloudWatch · GitHub ActionsRead the case study →
TICKET 006CASE STUDY

The Jenkins Job That Took 6 Tries

Provisioned Jenkins on EC2 with Terraform and documented all six failed attempts: AMI, Java version, wget, GPG order, nested heredocs and key rotation.

Terraform · AWS EC2 · Jenkins · Amazon Linux 2023 · S3Read the case study →
TICKET 005CASE STUDY

WordPress on a Three-Node Docker Swarm

A three-node Docker Swarm on EC2 running WordPress with three replicas and MySQL on separate overlay networks, deployed from the AWS CLI.

Docker Swarm · AWS EC2 · WordPress · MySQL · Overlay networkingRead the case study →
TICKET 004CASE STUDY

Serverless Text-to-Speech Pipeline

Amazon Polly text-to-speech driven by GitHub Actions: pull requests publish a beta audio file and merges to main publish production audio in S3.

AWS Polly · GitHub Actions · S3 · Python · boto3Read the case study →
TICKET 003CASE STUDY

Docker Swarm Self-Healing Container Infrastructure

A three-node Docker Swarm on EC2 running a global service, so every node runs a container and Swarm replaces any that go missing.

Docker Swarm · AWS EC2 · Overlay networking · UbuntuRead the case study →
TICKET 002CASE STUDY

Linux Security Audit Script

A Bash audit script on a systemd timer that flagged three real issues on my own machine on its first run, all fixed in about ten minutes.

Bash · Linux · systemdRead the case study →
TICKET 001CASE STUDY

Linux Process Sentinel

A Bash watchdog run by systemd that polls a service every 10 seconds, restarts it when it fails and logs every event.

Bash · Linux · systemdRead the case study →