Edges

Edges

Edges are the connections between nodes that define how data flows through your trading strategy. They represent the pathways along which information travels from one component to another.

What is an Edge?

An edge connects:

• A source node’s output handle (where data comes from)
• A target node’s input handle (where data goes to)

When data is produced by the source node, it flows along the edge to the target node’s input.

Edge Properties

Source

The node and output handle where data originates:

{
  "node": "ticker_btc",
  "handle": "price"
}

Target

The node and input handle where data is delivered:

{
  "node": "rsi_indicator",
  "handle": "value"
}

Data Type

The type of data flowing through the edge (must match source output and target input types).

State

The current state of the edge:

• Idle: No data flowing
• Active: Data is flowing
• Paused: Temporarily stopped
• Failed: Error in data transmission

Creating Edges

Visual Creation

1. Click on a source node’s output handle
2. Drag to the target node’s input handle
3. Release to create the edge

Type Validation

GrailHub validates that:

• Source output type matches target input type
• No circular dependencies are created
• Target input isn’t already connected (unless it accepts multiple connections)

Edge Behavior

Data Flow

When a source node produces output:

1. Data is sent along all connected edges
2. Target nodes receive the data on their inputs
3. Target nodes process the data
4. The cycle continues

Buffering

Edges can buffer data if the target node is busy:

• Default buffer size: 100 items
• Configurable per edge
• Prevents data loss during processing

Backpressure

If a target node can’t keep up:

• The buffer fills up
• Source node may slow down or pause
• Prevents memory overflow

Edge Types

Data Edges

Standard edges that carry data between nodes:

[Price Feed] --price--> [RSI Indicator]

Control Edges

Edges that control flow execution:

[Condition] --true--> [Execute Trade]
[Condition] --false--> [Log Event]

Event Edges

Edges that carry event notifications:

[Timer] --tick--> [Fetch Data]

Edge Visualization

Visual Indicators

Edges display visual cues:

• Solid line: Active connection
• Dashed line: Inactive connection
• Animated dots: Data flowing
• Red color: Error state
• Gray color: Idle state

Edge Labels

Edges can display labels showing:

• Data type
• Current value
• Flow rate

Managing Edges

Selecting Edges

Click on an edge to select it and view its properties.

Deleting Edges

Select an edge and press Delete or click the delete icon.

Reconnecting Edges

1. Click on the edge’s target end
2. Drag to a new target handle
3. Release to reconnect

Edge Patterns

Fan-Out

One source connected to multiple targets:

              /--> [Strategy A]
[Price Feed] ---> [Strategy B]
              \--> [Strategy C]

All strategies receive the same price data.

Fan-In

Multiple sources connected to one target:

[Source A] --\
[Source B] ----> [Aggregator]
[Source C] --/

The aggregator combines data from multiple sources.

Pipeline

Sequential processing:

[A] --> [B] --> [C] --> [D]

Data flows through each node in sequence.

Conditional

Branching based on conditions:

              /-true--> [Action A]
[Condition] --
              \-false-> [Action B]

Different actions based on condition result.

Data Transformation

Type Conversion

Some edges automatically convert data types:

[Number Output] --auto-convert--> [String Input]

Data Mapping

Map complex data structures:

[Object Output] --extract field--> [Specific Input]

Edge Monitoring

Flow Rate

Monitor how much data flows through an edge:

• Messages per second
• Bytes per second
• Average message size

Latency

Track how long data takes to traverse an edge:

• Average latency
• Maximum latency
• Latency distribution

Errors

Monitor edge errors:

• Type mismatches
• Buffer overflows
• Connection failures

Best Practices

Keep Flows Simple

Avoid creating overly complex edge patterns. Break into smaller flows if needed.

Use Descriptive Labels

Label edges to clarify what data they carry.

Monitor Buffer Sizes

Watch for buffer overflows indicating performance issues.

Handle Errors

Always connect error outputs to error handling nodes.

Avoid Circular Dependencies

Circular edges can cause infinite loops and deadlocks.

Test Edge Cases

Test what happens when:

• No data flows
• Too much data flows
• Invalid data is sent

Advanced Features

Conditional Edges

Edges that only activate under certain conditions.

Filtered Edges

Edges that filter data based on criteria.

Transformed Edges

Edges that transform data as it flows.

Prioritized Edges

Edges with different priorities for data delivery.

Debugging Edges

Inspecting Data

Click on an edge to see:

• Current data value
• Data history
• Flow statistics

Breakpoints

Set breakpoints on edges to pause when data flows.

Logging

Enable edge logging to track all data flowing through.

Next Steps

• Build Your First Flow
• Explore Advanced Patterns
• Optimize Performance