Shucked
CLIENT
ARTHOUSE Interactive
TECHNOLOGIES
WordPress, nginx, PHP-FPM, MariaDB, Redis, DigitalOcean, Cloudflare, Git, certbot
YEAR
2026
EXTERNAL URL
A zero-downtime hosting migration that lifted the Tony-winning musical Shucked off WP Engine onto a dedicated, self-managed DigitalOcean droplet — eliminating the production’s recurring hosting cost ahead of its national tour. The engagement wrapped a proper git and staging workflow around a site that had previously only been edited live in the WordPress admin.
Infrastructure, migration, and DevOps for ARTHOUSE Interactive — delivered with a rollback plan and an operations runbook.
Background
A Broadway show heading out on tour needed to shed a hosting bill it no longer wanted to carry.
Shucked, having closed on Broadway and preparing for a non-equity national tour, was still paying for premium managed hosting on WP Engine for a site that no longer needed it. ARTHOUSE Interactive brought me in to move the property onto infrastructure they controlled — cutting the ongoing cost while keeping the live site fully available throughout. Complicating the job: the theme had only ever been edited directly in the production WordPress admin, with no version control and no staging environment to test against.
The constraint that shaped everything: a public, ticket-selling site could not go dark — and the DNS cutover could not leave even a momentary TLS gap.
Implementation
Rebuild the stack on a droplet, add the workflow that was missing, then cut over with no gap.
I stood the site up on a self-managed DigitalOcean droplet — nginx, PHP-FPM, MariaDB, and a Redis object cache — and migrated the full database and uploads across from WP Engine. Around the theme I introduced the version control and deploy workflow it had never had: a bare git repository with a post-receive hook, and a branch model that maps develop to a staging environment and main to production, so changes could finally be tested before they went live.
The cutover was the delicate part. I pre-issued the TLS certificate for the domain on the new droplet before touching DNS, verified the site served correctly over HTTPS behind Cloudflare, and only then moved the DNS — eliminating any window where the site could be reachable without a valid certificate. A documented rollback path meant the change could be reversed at any point during the switch.
Production-readiness hardening and a written operations runbook rounded out the handoff, so the team could run and deploy the site without me.
Results
Same site, no downtime, no more managed-hosting bill — and a real deploy workflow.
The site moved to the self-managed droplet with no downtime and no TLS gap during the cutover, and the recurring WP Engine cost went away entirely. What had been a live-edit-only site now has version control, a staging environment, and push-to-deploy — so future changes are testable and reversible instead of applied straight to production.
Handed off with a rollback plan and an operations runbook, leaving the production self-sufficient on infrastructure it now owns.
