Reference
The reference describes how oncebox works and every setting it exposes.
| Page | What it covers |
|---|---|
| Core Concepts | Architecture, event lifecycle, batch capture, fencing, the database clock, background jobs |
| Delivery Guarantees | At-least-once, effectively-once and at-most-once, duplicates, ordering, limitations |
| Publishing Events | OutboxPublisher, @OutboxPublish, transactions, serialization, event ids |
| Processing & Resilience | State machine, retries and backoff, stuck event recovery, cleanup |
| Dead Letter Queue | DLQ state machine, transfer semantics, OutboxDlqHandler, REST API |
| Inbox | Idempotent consumers, choosing between them, cache, deserialization |
| Messaging Integration | Kafka and RabbitMQ senders, headers, routing, broker settings on both sides |
| Persistence | Databases, dialect modules, table schemas, MySQL isolation level |
| Schema Migrations | Creating and upgrading the tables with Flyway, Liquibase or at startup |
| Polling | Fixed and adaptive strategies, thread pool sizing, distributed lock |
| Observability | Metrics and Grafana dashboards |
| Configuration | Every property with its default |
| Configuration Recipes | Ready configurations: several event types, RabbitMQ, production, inbox, both sides, local development |