Blog

Engineering notes on Postgres and sovereign data.

Long-form pieces, written by the people building Lampion. Architecture, SQL schemas, production trade-offs — no press releases.

§ 01 — Latest article
Tutorial · 15 min read

Testing on production data, with the personal data masked

Fixtures lie, and a production dump on a developer’s laptop is processing personal data. A middle way: on every pull request, a copy-on-write fork of production, dynamic masking applied to the branch, your test suite running against real volumes while reading only fake values, and everything gone when the PR closes. The real data stays in the branch: this is not anonymisation, it is reduced exposure.

Read the article →
§ 02 — All articles
Guide 11 min read

Building a sovereign B2B SaaS with Lampion

Where to run the database, the API, the files and the logs when your customers are French companies — and how to keep each customer’s data away from the others. A six-step start, commands that follow on from each other, and the answers to the security questionnaire that arrives before the purchase order.
§ 03 — What we write about
Architecture
Multi-tenant data models, isolation, storage/compute separation, and what it changes in production.
Postgres
Row-level security, pooling, indexes, extensions, query plans — Postgres as we actually run it.
Workflow
Database branching, preview environments, migrations, CI/CD against real data.
Sovereignty
French hosting, GDPR, DPA, data residency — the concrete constraints, not the slogans.

A topic you’d like
to read about here?

Suggest a topic [email protected]