flowboard

Architecture in practice

Serverless processing

Separate the event source, stateless processor and durable output.

Start from template Download JSON

This opens a separate browser draft; your main draft remains unchanged.

Readable source

The labels explain the system to a reader. Stable identifiers connect the nodes to the flows. Keep a copy of this source, then adjust it to match your own architecture.

# Serverless processing
- Upload event {#upload type=service x=80 y=100}
- Process upload {#processor type=lambda x=400 y=100}
- Result store {#result type=database x=720 y=100}
## Flows
- upload -> processor {#trigger label="object created"}
- processor -> result {#write label="persist result"}

Make one useful change.

Rename a component, explain a connection or add a note about a missing assumption. Export the result as Markdown, JSON or YAML to keep an editable copy.

Learn the import and export workflow