Reliable delivery
At-least-once delivery with configurable retries and fallback handling. Consumers should be idempotent.
Watch · 58 seconds
Every event matters.
Meet Namastack Outbox and see how reliable event delivery fits into Spring Boot.
Transactional outbox
Keep database state and outgoing work consistent.
A database update and a broker publish cannot share one local transaction. Namastack stores the outgoing record beside the business change and processes it after commit.
01
Schedule atomically. Save the outbox record in the business transaction.
02
Process asynchronously. Partition-aware workers claim committed work.
03
Recover explicitly. Retries and fallback handlers deal with delivery failures.
Architecture flow
How it works
- Stage 01
Application transaction
Business data and the outbox record are persisted in one local transaction.
@Transactional - Stage 02
Commit
The committed outbox record becomes durable work ready for processing.
Durable state - Stage 03
Partitioned processing
Workers poll committed records and coordinate across application instances.
Partition aware - Stage 04
Handler execution
The matching handler runs while records with the same key stay ordered.
Per-key ordering - Stage 05
Completion or retry
Success completes the record; failures follow the configured retry strategy.
At least once
OrderService.ktKotlin · Spring transaction
// Business state and outgoing work commit together
@Transactional
fun createOrder(command: CreateOrderCommand) {
val order = orderRepository.save(Order.create(command))
outbox.schedule(
payload = OrderCreatedEvent(order.id, order.customerId),
key = "order-${order.id}"
)
}Product capabilities
One focused library, from persistence to operations
Choose the modules your application needs and keep the processing model inside Spring Boot.
Strict ordering
Records that share a key retain their processing order across polling cycles.
Horizontal scaling
Partition-aware processing coordinates work safely across multiple application instances.
Flexible persistence
Choose JPA, JDBC, or MongoDB to fit the persistence model already used by the application.
Messaging integrations
Publish through Kafka, RabbitMQ, and Amazon SNS integrations, or implement a custom handler.
Observability
Micrometer metrics, distributed tracing, and context propagation cover the asynchronous boundary.
Understand the delivery contract
At-least-once delivery can produce duplicates. The guarantees guide explains ordering, retries, transactions, and consumer responsibilities.
Spring Modulith integration
Use Namastack as the outbox behind Spring Modulith events.
Namastack Outbox is an independent library. Spring Modulith applications can use the official starter to externalize events through Namastack's persistence, processing, retry, ordering, and observability model.
Read the integration guideSpring Modulith
spring-modulith-starter-namastackNamastack OutboxJPA · JDBC · MongoDBStart with a working outbox.
Install a persistence starter, schedule your first record, and follow the operational model in the documentation.
Support the work
Namastack Outbox is independently developed and maintained as open-source software. Sponsorship supports maintenance, documentation, integrations, and long-term stability.