workflows

name	workflows
description	Build Glide workflows with automation, loops, conditions, queries, and background processes. Use when creating automations, processing data in bulk, building scheduled tasks, or integrating with external services.

Glide Workflows

Workflows are Glide's automation engine for background processes, data transformations, and integrations. They run server-side without user interaction.

What Workflows Do

Workflows automate tasks that would otherwise require manual work:

• Send notifications when data changes
• Process data in bulk using loops
• Connect to external APIs and services
• Run scheduled reports or cleanup tasks
• Transform and move data between tables
• Multi-step business processes

When to Use Workflows

• Event-driven automation: "When a new order is created, send email confirmation"
• Scheduled tasks: "Every Monday at 9am, generate weekly report"
• Bulk processing: "For each pending invoice, calculate total and send reminder"
• External integrations: "Sync data from external API to Glide table"
• Data transformation: "When spreadsheet is uploaded, parse and distribute to multiple tables"

Accessing Workflows

In Glide Builder:

1. Click Workflows tab in the top navigation
2. Click "+ New Workflow" to create
3. Each workflow has:
   • Name - descriptive label
   • Trigger - what starts the workflow
   • Steps - nodes that execute in sequence

Triggers

Triggers determine when a workflow runs.

Event Triggers

Row Added

• Runs when a new row is created in a table
• Access the new row's data in workflow steps
• Use case: Send welcome email when user signs up

Row Updated

• Runs when any column in a row changes
• Can filter to specific column changes
• Use case: Notify manager when order status changes to "Shipped"

Row Deleted

• Runs when a row is removed from a table
• Can access the deleted row's data
• Use case: Log deletions for audit trail

Webhook

• Runs when external service sends HTTP request to webhook URL
• Receives POST data as workflow input
• Use case: Stripe payment webhook triggers order fulfillment

Scheduled Triggers

Schedule

• Runs on a recurring schedule (hourly, daily, weekly, monthly)
• Set specific time and timezone
• Use case: Every Sunday at midnight, archive completed tasks

Custom Schedule (Cron)

• Advanced scheduling using cron expressions
• Full control over timing
• Use case: 0 9 * * 1-5 = 9am weekdays only

Manual Triggers

Button Action

• Triggered by user clicking a button in the app
• Useful for on-demand processing
• Use case: "Generate Report" button runs export workflow

Form Submission

• Triggered when form is submitted
• Access form data in workflow
• Use case: After contact form submission, create task and notify team

Workflow Nodes

Nodes are the building blocks of workflows. They execute sequentially from top to bottom.

Query Node

Fetch data from Glide tables to use in the workflow.

Configuration:

• Table: Which table to query
• Filters: Optional conditions to narrow results (e.g., Status = "Active")
• Limit: Max rows to return (useful for testing)
• Columns: Which columns to include (all by default)

Output: Returns an array of row objects: [{col1: val1, col2: val2}, ...]

Common uses:

• Get all pending orders to process
• Find users matching criteria
• Fetch related data for processing

Example pattern:

Query Node → Get Orders where Status = "Pending"
  ↓
Loop Node → Process each order

Loop Node

Iterate over an array, running child nodes for each item.

Configuration:

• Array: The data to loop over (typically from Query Node)
• Item Name: Reference name for current item (e.g., "order")
• Max Iterations: Safety limit to prevent runaway loops

How it works:

• Takes an array as input
• Runs all child nodes once for each array item
• Inside loop, access current item data via item reference
• Child nodes are indented under the loop

Common uses:

• Process each row from a Query
• Send email to each user in a list
• Create multiple rows based on template

Example:

Query → Get pending tasks
  ↓
Loop over tasks (item: "task")
  ├─ Condition → If task.priority = "High"
  │    └─ Send Email → Notify manager
  └─ Update Row → Mark task as "Processed"

Condition Node

Add if/then logic to control workflow execution.

Configuration:

• Condition: Expression that evaluates to true/false
• Then steps: Nodes that run if condition is true
• Else steps: Nodes that run if condition is false

Conditions can check:

• Field values: status = "approved"
• Comparisons: total > 100
• Existence: email is not empty
• Combinations: priority = "High" AND dueDate < today

Common uses:

• Skip processing if data doesn't meet criteria
• Branch workflow based on values
• Prevent errors (e.g., only send email if email field isn't empty)
• Implement business rules

Example:

Condition → If order.total > 1000
  THEN:
    └─ Send Email → Notify VIP sales team
  ELSE:
    └─ Send Email → Notify standard sales team

Set Column Node

Extract or transform data, often used to prepare data for other nodes.

Configuration:

• Input: The source data (object or array)
• Column/Field: Which field to extract
• Operation: Extract, Transform, Calculate

Critical for Query Result Processing:

When you query a table, you get an array of objects:

[
  {email: "[e-mail]", name: "Alice", status: "Active"},
  {email: "[e-mail]", name: "Bob", status: "Active"}
]

To extract just the emails into a simple array:

Query Node → Get users where status = "Active"
  ↓
Set Column Node → Extract "email" column
  ↓
Result: ["[e-mail]", "[e-mail]"]

Why this matters:

• Many nodes expect simple arrays, not arrays of objects
• Email nodes want string array of addresses
• API calls may need specific field formats

Common pattern: Query → Extract → Loop

Query → Get customers
  ↓
Set Column → Extract customer.email
  ↓
Loop → For each email
    └─ Send Email → email

Create Row Node

Add a new row to a table.

Configuration:

• Table: Target table
• Column Values: Data for each column

Common uses:

• Log workflow execution
• Create derived records
• Duplicate rows with modifications

Example:

Trigger: Order created
  ↓
Create Row → Add row to "Order Logs" table
  - Order ID: {trigger.orderID}
  - Timestamp: {now}
  - Status: "Processed"

Update Row Node

Modify an existing row in a table.

Configuration:

• Table: Which table
• Row: Which row to update (by ID or from loop item)
• Column Values: New values to set

Common uses:

• Mark records as processed
• Update status after action
• Calculate and store derived values

Example:

Loop over pending orders (item: "order")
  ├─ API Call → Get shipping cost
  └─ Update Row → order
      - Shipping Cost: {API response}
      - Status: "Quoted"

Delete Row Node

Remove a row from a table.

Configuration:

• Table: Which table
• Row: Which row to delete (by ID or from loop item)

Common uses:

• Clean up temporary data
• Archive old records
• Remove duplicates

Send Email Node

Send email notifications.

Configuration:

• To: Recipient email(s) - can be array from Set Column
• Subject: Email subject line
• Body: Email content (plain text or HTML)
• From: Optional custom sender name

Works with arrays:

Set Column → Extract emails from query result
  ↓
Send Email → To: {extracted emails array}

Common uses:

• Notify users of status changes
• Send bulk announcements
• Alert admins of errors

API Request Node

Call external web services via HTTP.

Configuration:

• Method: GET, POST, PUT, DELETE
• URL: API endpoint
• Headers: Authentication, Content-Type, etc.
• Body: Request payload (for POST/PUT)

Response handling:

• Store response in variable
• Use response data in subsequent nodes
• Parse JSON responses

Common uses:

• Fetch data from external API
• Send data to external system
• Trigger webhooks in other services
• Integration with Stripe, SendGrid, etc.

Example:

Trigger: Form submitted
  ↓
API Request → POST to Slack webhook
  - URL: https://hooks.slack.com/...
  - Body: {"text": "New form submission from {name}"}

JavaScript Node

Run custom JavaScript code for complex logic.

Use cases:

• Advanced data transformation
• Complex calculations
• Custom parsing logic
• Validation rules

Has access to:

• Workflow variables
• Input data
• JavaScript standard library

Example:

// Parse and validate phone number
const phone = input.phoneNumber;
const cleaned = phone.replace(/\D/g, '');
return cleaned.length === 10 ? cleaned : null;

Wait/Delay Node

Pause workflow execution.

Configuration:

• Duration: How long to wait (seconds, minutes, hours, days)

Common uses:

• Rate limiting (wait between API calls)
• Scheduled follow-ups (wait 3 days, then send reminder)
• Throttling bulk operations

Example:

Loop over 1000 contacts
  ├─ Send Email → contact
  └─ Wait → 1 second (prevent rate limiting)

Advanced Patterns

Query → Extract Column → Loop Pattern

The most common workflow pattern for bulk processing:

Query Node → Get all users where status = "Active"
  Output: [{id: 1, email: "[e-mail]", name: "Alice"}, ...]
  ↓
Set Column Node → Extract "email" field
  Output: ["[e-mail]", "[e-mail]", "[e-mail]"]
  ↓
Loop Node → For each email
  Item: "email"
  ↓
  Send Email Node → To: {email}
    Subject: "Weekly Newsletter"

Why extract the column?

• Query returns full row objects
• Most action nodes (Email, API) need simple values
• Extracting creates a clean array of just the values you need

Nested Loops for Multi-Level Processing

Process hierarchical data by nesting loops:

Query → Get all projects
  ↓
Loop over projects (item: "project")
  ↓
  Query → Get tasks where projectID = {project.id}
    ↓
  Loop over tasks (item: "task")
    ↓
    Condition → If task.status = "Overdue"
      ↓
      Send Email → Notify {task.assignee}

Error Handling with Conditions

Prevent workflow failures by validating data:

Trigger: Form submission
  ↓
Condition → If email is not empty
  THEN:
    └─ Send Email → {email}
  ELSE:
    └─ Create Row → Error log
        - Message: "Missing email"
        - Form ID: {form.id}

Batch Processing with Limits

Test with small batches before running on full dataset:

Query → Get pending invoices
  - Limit: 10 (for testing)
  ↓
Loop over invoices
  ↓
  API Request → Send to accounting system
  Wait → 500ms
  Update Row → Mark as "Synced"

Once tested, remove or increase the limit.

Conditional Branching by Type

Handle different record types in one workflow:

Trigger: Row added to "Notifications" table
  ↓
Condition → If notificationType = "Email"
  THEN:
    └─ Send Email
  ELSE:
    Condition → If notificationType = "SMS"
      THEN:
        └─ API Request → Twilio SMS
      ELSE:
        └─ Create Row → Error log

Data Aggregation

Collect results from loop and store:

Query → Get all orders from this month
  ↓
JavaScript Node:
  // Calculate total revenue
  let total = 0;
  for (const order of input.orders) {
    total += order.amount;
  }
  return {total, count: input.orders.length};
  ↓
Update Row → Monthly Stats table
  - Revenue: {total}
  - Order Count: {count}

Testing Workflows

Test Mode

Before enabling a workflow for production:

1. Set Row Limits: Add limit to Query nodes (e.g., 5 rows)
2. Use Test Data: Create test rows to trigger workflow
3. Check Execution Logs: View workflow runs in history
4. Verify Output: Check that data changes are correct

Execution History

In Workflows tab:

• Click on a workflow to see execution history
• Each run shows:
  • Timestamp
  • Trigger source
  • Status (Success/Failed)
  • Execution time
  • Error messages (if failed)

Debugging Failed Workflows

If a workflow fails:

1. Check error message in execution history
2. Common issues:
   • Missing required fields
   • API authentication failed
   • Invalid data format
   • Timeout (workflow took too long)
   • Rate limit exceeded
3. Add logging: Use Create Row to log progress
4. Add conditions: Validate data before risky operations
5. Test incrementally: Enable one node at a time

Logging Pattern for Debugging

Query → Get pending items
  ↓
Create Row → Workflow Log
  - Message: "Found {count} items to process"
  - Timestamp: {now}
  ↓
Loop over items
  ↓
  [Process item]
  ↓
  Create Row → Workflow Log
    - Message: "Processed item {item.id}"

Performance & Limits

Quota Costs

Workflows consume update quota:

• Each row read: 0.001 updates
• Each row write: 0.01 updates
• API calls: May have their own limits

Timeouts

• Workflows timeout after 5 minutes
• For long-running tasks, use Wait nodes strategically
• Consider splitting into multiple workflows

Rate Limiting

When calling external APIs:

• Add Wait nodes between calls
• Batch requests when possible
• Handle rate limit errors gracefully

Best Practices

1. Test with limits: Use Query limits during development
2. Add conditions: Validate data before processing
3. Log progress: Create audit trail for debugging
4. Handle errors: Use conditions to catch edge cases
5. Batch operations: Process in chunks if dataset is large
6. Use indexes: Filter queries to reduce data scanned
7. Avoid infinite loops: Always set max iterations on loops
8. Monitor quota: Track update usage in large workflows

Common Use Cases

Send Welcome Email on Signup

Trigger: Row added to Users table
  ↓
Condition → If user.email is not empty
  THEN:
    Send Email
      - To: {user.email}
      - Subject: "Welcome to {appName}!"
      - Body: "Hi {user.name}, thanks for joining..."

Weekly Digest Email

Trigger: Schedule (Every Monday 9am)
  ↓
Query → Get users where emailPreference = "Weekly"
  ↓
Set Column → Extract email
  ↓
Query → Get new content from last 7 days
  ↓
JavaScript → Format content into HTML
  ↓
Send Email
  - To: {user emails array}
  - Subject: "Your Weekly Digest"
  - Body: {formatted HTML}

Sync External Data

Trigger: Schedule (Every hour)
  ↓
API Request → GET from external API
  - URL: https://api.example.com/products
  ↓
Loop over API response items
  ↓
  Query → Check if product exists (by externalID)
  ↓
  Condition → If product exists
    THEN:
      Update Row → Update price, stock
    ELSE:
      Create Row → Add new product

Order Fulfillment

Trigger: Order status updated to "Paid"
  ↓
Condition → If order.total > 0
  THEN:
    API Request → POST to shipping service
      - Create shipping label
    ↓
    Update Row → order
      - Tracking Number: {API response}
      - Status: "Shipped"
    ↓
    Send Email → order.customerEmail
      - Subject: "Order shipped!"
      - Body: "Track your package: {trackingNumber}"

Data Cleanup

Trigger: Schedule (Daily at 2am)
  ↓
Query → Get rows where createdAt < 90 days ago AND status = "Temporary"
  ↓
Loop over old rows
  ↓
  Delete Row → item
  ↓
Create Row → Audit Log
  - Message: "Deleted {count} old temporary records"
  - Timestamp: {now}

Workflows vs. Computed Columns

Know when to use each:

Feature	Computed Column	Workflow
Timing	Instant (during app use)	Background (delayed)
Use case	Calculations, lookups	Automation, notifications
Scope	Single row	Multiple rows, external systems
Examples	Total = Price × Qty	Send email when status changes

Use computed columns when:

• Calculation based on current row
• Need instant result in UI
• Simple logic (math, if-then, lookup)

Use workflows when:

• Triggering external actions
• Processing multiple rows
• Sending notifications
• Scheduled tasks
• Complex multi-step processes

Summary

Workflows are powerful for:

• ✅ Event-driven automation
• ✅ Scheduled tasks
• ✅ Bulk data processing
• ✅ External integrations
• ✅ Multi-step business processes

Key patterns:

• Query → Extract → Loop for bulk processing
• Condition nodes for validation and branching
• Set Column to convert query results to arrays
• Wait nodes for rate limiting
• Logging for debugging

Always test with limits before running on production data.