FAQ
Does the AI score my work?
No. Scoring is fully deterministic and offline. The same fix always earns the same Detect / Fix / Trap result. The AI is used only to generate hints, never to grade you.
Do I need Docker?
Yes. rillence postgres runs a real PostgreSQL instance in Docker for every scenario, so Docker must be installed and running before you start.
Is it safe to break the database?
Completely. Each scenario runs in a disposable sandbox. Drop tables, corrupt data, or wedge the server however you like — nothing leaves the container. Run reset to roll back to the starting state at any time.
Does scoring need internet?
No. Scoring runs locally against your sandbox, so you can grade your work fully offline. Only AI hints require a connection.
How many scenarios are there?
There are scenarios across the tracks, ranging from everyday operational mishaps to deep, subtle failure modes.
Can I retry a scenario?
Yes. Run reset to restore the scenario's initial state and try a different approach. You can retry as many times as you want.
Do hints lower my score?
No. Hints are tracked but not penalized. Your Detect / Fix / Trap score reflects only the state of the database when you run score, so reach for hint whenever you're stuck.
Does it touch production?
No. rillence postgres only ever talks to its local Docker sandbox. It never connects to, reads from, or modifies any production system.