WRITING
Why we built what we built, how we think about reliability when nobody's watching, and the technical ideas behind Acheron Vault, Blackglass, and Charon Gate.
Essays on backup testing, restore evidence, and why scheduled jobs are not verification. Acheron Vault →
Running pg_dump on a schedule is not a backup test — it is a backup creation process. A real test performs an actual restore against the backup file, checks integrity, records timing, and produces evidence. Most teams have never done one outside a crisis.
Scheduled checks run at the wrong time. Your highest-risk windows open immediately after something changes — a migration, a deployment, a configuration update. Clocks don't know when those happen. Events do.
How inbound webhooks fail in production, and what durable ingest actually requires. Charon Gate →
Every webhook integration eventually goes wrong in the same five ways. We built Charon Gate because gluing the same retry-and-DLQ stack into every service stopped being a reasonable use of an engineering team's time.
HMAC, immediate ACK, jittered backoff, dead-letter queue, replay. The five primitives that decide whether your webhook handler is a production system or a quietly-failing background task.
Baselines, drift, and the operational picture Linux teams actually need. Blackglass →
Drift is rarely dramatic. It is the package installed during an incident, the SSH setting left open after a test, the cron job nobody remembers. Here is what it looks like on real Linux fleets.
A baseline is "the picture of this server we agreed was fine". Everything Blackglass does, including the things that look fancy in a screenshot, is downstream of that idea.
More posts as we ship things worth writing about. No content calendar.