Flowintent / proof
Loading docs
Operate the Docker stack, verify health checks, understand service boundaries, and prepare production deployment.
Deployment documentation is for operators who need to run the application stack reliably. The main services are web, database, Redis, Kafka, quantum runner, hardware poller, and Nginx.
Use this page as a product-level orientation. Detailed environment and production operations remain in the repository ops docs.
This page belongs to the Security & Ops section and should be read with the related pages listed at the end.
Use screenshots and notes to explain product behavior without exposing private credentials, admin state, or customer data.
The web service serves the Next.js application and API routes. Postgres stores product data. Redis supports fast state and queues. Kafka carries event and notification workflows. The quantum runner handles quantum execution. The hardware poller reconciles provider jobs and notifications out of band. Nginx exposes the public HTTP surface.
Before production rollout, verify each dependency from the health endpoint and from Docker health state.
The /api/health endpoint always reports database and runner status, and adds Redis and Kafka only when those are configured (REDIS_URL, KAFKA_BROKERS). The database always gates overall health; the runner gates only when simulations are routed to the Python service (QUANTUM_BACKEND=python or QUANTUM_RUNNER_REQUIRED=1); Kafka gates only when QW_KAFKA_REQUIRED=1. Redis is reported for visibility but does not gate.
When a gating check fails, the endpoint answers 503. Debug that dependency directly before chasing UI symptoms.
Use observability surfaces to monitor provider readiness, workflow readiness, hardware setup, and operational signals.
Understand provider credential boundaries, reviewer-safe shares, role isolation, and safe documentation capture.
Definitions for common QFlow Studio terms: workflow, block, provider, route, run, trace, evidence, academy, and teacher source.