REFERENCE ARCHITECTURE
Event-driven integration
Asynchronous workflows with Conduit (Kafka), Lane (RabbitMQ), and Beacon (NATS) — decouple producers from consumers and scale workers independently.
IntermediatePlan: Business+Resilience: multi-zona
Diagram
Use cases
- Order fulfillment
- Webhook fan-out
- Legacy system integration
Components
| Layer | Role | Product |
|---|---|---|
| Ingress | Public API entry | Gateway + Compute |
| Event log | Durable ordered streams | Conduit |
| Job queues | Task dispatch + dead-letter | Lane |
| Pub/sub | Lightweight notifications | Beacon |
| Workers | Idempotent consumers | Compute + Autoscale |
| State | Checkpoints and audit trail | Datastore |
| Operations | Queue lag and DLQ alerts | Insight |
Products
Conduit
View product docs →Beacon
View product docs →Compute
View product docs →Datastore
View product docs →Insight
View product docs →When to use
- Order pipelines, webhook fan-out, audit event streams.
- Peak traffic buffering (Black Friday, batch imports).
- Integrating legacy systems without synchronous coupling.
When to avoid
- Simple CRUD with immediate read-your-writes — synchronous API is simpler.
- Teams without idempotent consumer design discipline.
Design notes
Pick the right bus
Conduit for durable event logs and replay. Lane for job queues with ACK. Beacon for lightweight pub/sub notifications.
Idempotent workers
Design Compute workers to handle duplicate messages. Store processing checkpoints in Datastore.
Dead-letter strategy
Route failed Lane messages to a dedicated queue and alert via Insight.
Blueprint
Use this JSON as a starting point when creating a project via the Tesserra API or console. Replace image URLs, domains, and resource references with your values.
json
{
"nome": "Order Pipeline",
"modo": "avancado",
"recursos": [
{
"tipo": "compute",
"nome": "orders-api",
"config": {
"subtipo": "api",
"porta": 3000,
"replicas_min": 2
}
},
{
"tipo": "conduit",
"nome": "order-events",
"config": {
"particoes": 6,
"retencao_horas": 168
}
},
{
"tipo": "lane",
"nome": "fulfillment-jobs",
"config": {
"prefetch": 10,
"duravel": true
}
},
{
"tipo": "beacon",
"nome": "notifications",
"config": {
"jetstream": true
}
},
{
"tipo": "compute",
"nome": "fulfillment-worker",
"config": {
"subtipo": "worker",
"replicas_min": 1,
"replicas_max": 8
}
},
{
"tipo": "datastore",
"nome": "pipeline-db",
"config": {
"engine": "postgres",
"tamanho_gb": 30
}
},
{
"tipo": "insight",
"nome": "pipeline-insight",
"config": {
"retencao_dias": 30,
"alertas": true
}
}
]
}Related
Microservices on private Fabric
Multiple Compute services on an isolated Fabric network, public ingress via Gateway, per-service Datastores and Autoscale policies.
AI inference workload
Serve models via Compute (self-hosted) or Tesserra managed agents — cache with Datastore Redis, audit in Postgres, datasets in Archive.
Single-region web application
Production-ready pattern for SPAs, e-commerce, and public APIs — Compute, Datastore, Gateway, Sentinel, and Insight in one region.