April 11, 2026
Absurd name, serious drama
In Production
Runs big jobs with one database — fans cheer, purists cringe, and everyone dunks on the title
TLDR: Absurd is now running in production with a simple, one-database design and handy tools like a CLI and dashboard. Comments zeroed in on the mangled title while the classic split—simplicity lovers vs. workflow-engine purists—simmered, highlighting the hunger for fewer moving parts when running reliable long jobs.
Absurd, a tool that keeps long-running computer chores safe and restartable, just graduated to real-world use — and the crowd immediately got stuck on the headline. One early comment nitpicked the title (“Absurd In Production”), and the name itself turned into a punchline. Irony alert: calling it “Absurd in production” felt like both a flex and a threat to some readers.
Beyond the title drama, people were buzzing about the simplicity: no extra services, no complicated setup — just a single SQL file and thin language helpers. The author says it held up under pressure and added practical goodies: a command-line fixer-upper, a clean web dashboard, and the ability to pause, resume, and fetch results later. Translation: fewer moving parts, more sleep for on-call teams.
But the vibe split fast. The “one database to rule it all” camp loves that it lives fully inside Postgres — fewer bills, fewer headaches. The skeptics clutched pearls at using a database as a job runner, warning this road is paved with lost weekends and mystery deadlocks. Memes followed: “Absurdctl my life,” jokes about the title reading like a postmortem, and the eternal “just use a proper workflow engine” chorus. It’s equal parts admiration and anxiety — perfectly Absurd, really.
Key Points
- •Absurd is a Postgres-native durable execution system centered on a single SQL file with SDKs for TypeScript, Python, and experimental Go.
- •The team has run Absurd in production for five months, releasing multiple updates to harden reliability (claim handling, watchdogs, deadlock prevention, lease management, event race conditions).
- •New decomposed step APIs (beginStep()/completeStep()) enable inspecting step state before committing, aiding conditional logic and hook integrations.
- •Task result handling now allows spawning tasks and later fetching/awaiting outcomes, enabling parent-child workflow patterns and improved debugging.
- •New tooling includes the absurdctl CLI for operations and debugging and Habitat, a Go web dashboard for live monitoring; agent-focused integrations were also added.