BentoViz

Work in Progress, do not use unless for contributing.

A visual node editor for Bento stream processing workflows, built with LiteGraph.js.

Overview

BentoViz provides a graphical interface for designing, configuring, and deploying Bento stream processing pipelines. It allows you to:

• Visually compose input → processor → output pipelines
• Auto-generate configuration from the Bento JSON schema
• Deploy streams directly to a running Bento instance
• Save and load workflow graphs
• Export configurations in YAML or JSON format

Features

• Visual Node Editor: Drag-and-drop interface for building stream pipelines
• Schema-Driven: Automatically generates nodes from Bento's JSON schema (inputs, processors, outputs)
• Live Deployment: Deploy streams to Bento via REST API
• YAML Export: Generate clean YAML configurations
• Auto-Save: Graphs are automatically saved to browser storage
• Stream Management: List, deploy, and stop streams from the UI
• Special Processor Support: Visual editing for processors with nested processors (branch, retry, switch, group_by)

Prerequisites

• Go 1.21+
• A running Bento instance (for stream deployment)

Building

go install github.com/akhenakh/bentoviz@latest

Or download a binary.

Running

You need a running Bento: bento streams.

# Start BentoViz (default: port 8080, proxies to localhost:4195)
./bentoviz

# Custom port and Bento URL
./bentoviz -port 3000 -bento http://localhost:4195

# Show help
./bentoviz -h

Usage

1. Open the Editor: Navigate to http://localhost:8080 in your browser

2. Add Nodes:

   • Double-click or drag nodes from the left palette into the canvas
   • Nodes are organized by category: Inputs, Processors, Outputs

3. Connect Nodes:

   • Drag from an output port to an input port to create connections
   • Pipeline flow: Input → Processor(s) → Output

4. Configure Nodes:

   • Click on a node to see its properties
   • Click on text fields to edit values
   • For multiline fields (mapping, bloblang), a popup dialog appears

5. Deploy Stream:

   • Enter a Stream ID in the sidebar
   • Click "Deploy Stream" to push to Bento
   • View active streams in the Streams list

6. Export Configuration:

   • Click "Preview Config" to see the generated YAML
   • Click the copy button to copy to clipboard

7. Save/Load Graphs:

   • "Save" button downloads the current graph as JSON
   • "Load" button imports a previously saved graph

Special Processors

Some Bento processors have nested processor lists. BentoViz provides special handling for these:

Branch Processor

The branch processor executes processors on a mapped copy of the message:

• Connect processors to the processors output (right side)
• Configure request_map and result_map fields

- branch:
    request_map: 'root = {"id": this.id}'
    processors:
      - http:
          url: http://example.com/api
    result_map: 'root = this'

Retry Processor

The retry processor retries failed messages:

• Connect processors to the processors output (right side)
• Configure max_retries and parallel options

- retry:
    max_retries: 3
    parallel: false
    processors:
      - http:
          url: http://example.com/api

Switch Processor

The switch processor routes messages to different processor chains based on conditions:

• Click "+ Add Case" button to add cases
• Each case has a check condition field
• Connect processors to each case's output (right side)

- switch:
    - check: this.type == "foo"
      processors:
        - mapping: 'root.result = "foo"'
    - check: this.type == "bar"
      processors:
        - mapping: 'root.result = "bar"'

Group By Processor

The group_by processor groups messages and processes each group:

• Click "+ Add Group" button to add groups
• Each group has a check condition field
• Connect processors to each group's output (right side)

- group_by:
    - check: content().contains("foo")
      processors:
        - archive:
            format: tar
        - mapping: 'meta grouping = "foo"'

Label Field

All processors have an optional label field for uniquely identifying them in observability data (metrics, logs):

- label: my-processor
  mapping: |
    root = this

Development

Updating the Schema

1. Generate Bento schema:

   # inside bento repository
   go run ./cmd/tools/bento_docs_gen

2. Rebuild the binary:

   go build -o bentoviz .

Special Processors in Code

Special processors (branch, retry, switch, group_by) are handled differently:

• Branch/Retry: Use processorsAnchor flag; have a processors output port
• Switch/Group By: Use switchNode flag; have dynamic case/group outputs with buttons

API Endpoints

The Go server proxies the following endpoints to Bento:

Endpoint	Method	Description
/streams	GET	List all streams
/streams/{id}	POST	Create/update stream
/streams/{id}	DELETE	Delete stream

License

MIT License - see LICENSE file for details.