How Workflows Run

Learn what happens when your workflows execute and how to monitor their performance.

Workflow Execution Basics

What Happens When a Workflow Starts

Trigger Fires - Something starts your workflow
- Webhook receives data
- Schedule time arrives
- File gets uploaded
- Button gets clicked
Data Preparation - System gets ready
- Loads workflow definition
- Prepares variables and context
- Validates permissions
- Allocates resources
Step Execution - Actions happen in order
- Each step receives input data
- Performs its specific action
- Produces output for next step
- Reports success or failure
Completion - Workflow finishes
- Final results recorded
- Resources cleaned up
- Notifications sent
- Status updated

Execution Modes

Sequential Execution Steps run one after another:

Step 1 completes → Step 2 starts
Predictable order
Easy to follow and debug
Best for dependent operations

Parallel Execution Multiple steps run simultaneously:

Faster overall completion
Better resource utilization
Handles independent operations
Automatic coordination and joining

Conditional Execution Different paths based on decisions:

Dynamic routing
Business rule enforcement
Exception handling
Flexible process flow

Variable Processing

How Variables Work

Variables connect your workflow steps by passing data between them.

Variable Resolution When you use {{variable.name}}, the system:

Looks for the variable in step outputs
Checks global variables
Searches trigger data
Uses context information
Returns the value or reports error

Data Types Variables can hold different types of data:

Text - Names, descriptions, messages
Numbers - Amounts, quantities, calculations
Dates - Timestamps, schedules, deadlines
Lists - Collections of items to process
Objects - Complex data structures

Example Variable Usage

Email Subject: "Order {{order.id}} shipped to {{customer.name}}"
API URL: "https://api.example.com/customers/{{customer.id}}"
Condition: "{{invoice.amount}} > {{limits.approval_threshold}}"

Dynamic Data Flow

Step-to-Step Data Passing

Output from Step 1 becomes input for Step 2
Calculations and transformations happen automatically
Complex data structures maintained
Error handling preserves data integrity

Global Variable Access

Organization settings available everywhere
API keys and configuration values
Business rules and thresholds
Shared resources and connections

Error Handling & Recovery

Automatic Error Handling

Retry Logic When steps fail, the system automatically:

Waits a short time (1 second)
Retries the step
If still failing, waits longer (4 seconds)
Retries again with exponential backoff
After maximum attempts, escalates to manual review

Smart Retry Decisions The system knows which errors to retry:

Retry: Network timeouts, server busy, rate limits
Don't Retry: Invalid data, authentication errors, business rule violations

Manual Interventions

When Human Review is Needed

Data validation failures
External system unavailable
Business rule exceptions
Approval requirements

Intervention Process

Workflow pauses automatically
Notification sent to designated reviewers
Issue appears in dashboard
Reviewer can:
- Fix data and continue
- Retry the failed step
- Skip the problematic step
- Abort the workflow