Case Studies

A few examples of the sort of work we do. Names and some details are softened, but the shape of the problems – and the way we solved them – is exactly as it tends to appear in the wild.

The Haunted Lighthouse itself – eating our own cooking

From scattered experiments to a coherent infrastructure

Digital infrastructure & platforms

The Haunted Lighthouse Limited started life as many small technical outfits do: a collection of servers, domains and ideas, with a strong sense of direction but a slightly improvised set of tools.

The situation

• A self-hosted Mastodon instance, moving through several iterations.
• Git repositories scattered between local machines and third-party platforms.
• Backup scripts that worked, but only in the head of the person who wrote them.
• Multiple machines (desktop, laptop, servers) all playing a role with no single map.

What we did

• Standardised on Forgejo for private Git hosting and established a clear backup pipeline.
• Moved Mastodon to a more predictable, documented footing with monitoring and logging.
• Introduced MinIO for S3-compatible storage and folded it into backup and retention plans.
• Documented where everything lives, how it is backed up, and how to rebuild it.

The outcome

Instead of a collection of “clever one-offs”, Lighthouse now runs as a small but coherent system: Git, Mastodon, storage, backups and monitoring all know about each other. New services can be added without the whole thing feeling like a Jenga tower.

A small software team – from “just GitHub” to something more deliberate

Bringing structure to a growing codebase

Forgejo & CI/CD

A three-person software team had grown beyond “push everything to GitHub and hope”. They were concerned about lock-in, wanted more control over where their code and artefacts lived, and needed a cleaner path to production.

The situation

• Several critical projects spread across personal GitHub accounts.
• CI pipelines defined ad hoc, duplicated across repositories.
• No single place to store build artefacts or documentation.
• No clear backup or restore story beyond “GitHub has it, probably”.

What we did

• Migrated their active projects into a managed Forgejo organisation under Lighthouse.
• Set up standard CI/CD templates they could reuse and adapt.
• Provisioned a MinIO bucket for build artefacts and documentation bundles.
• Folded Forgejo into our existing encrypted backup regime with a simple restoration plan.

The outcome

The team kept their existing workflows while gaining control, traceability and a clear backup posture. Their code and artefacts now live in a space they can describe to clients and partners without hand-waving.

A solo consultant – getting out of the “too many services” trap

Making a modest estate manageable

Rationalisation & backups

An independent consultant had collected a sensible but slightly unplanned set of tools: shared hosting here, a project management SaaS there, cloud storage in several places, a handful of domains and a vague sense of unease.

The situation

• Multiple providers for email, storage and hosting.
• No clear answer to “what gets backed up and where?”.
• Old projects still live on public services out of habit.
• Concern about privacy and data locality for client material.

What we did

• Mapped every domain, service and provider into a single diagram.
• Retired a few unused services and consolidated others.
• Moved source code and internal documentation into Forgejo and S3 storage.
• Set up a simple backup pattern the consultant could run and understand themselves.

The outcome

Instead of a fog of logins and half-remembered subscriptions, the consultant now has a small, well-documented estate that is easy to explain and cheap to run. If something breaks, they know where to look and who to ask.

Something here feel familiar?

If you recognise parts of your own setup in any of these cases, we can probably help. The details will differ, but the underlying themes tend to repeat.

Talk about your own case