Skip to main content

Agent Execution

Learn how to run, monitor, and manage AI agents in Autonify. This guide covers manual execution, automatic triggers, scheduling, and configuration options for optimal agent performance.

VIDEO TUTORIAL⏱️ 2 min

📹 Agent Configuration & Execution Tutorial

Master AI agent execution in Autonify with this detailed tutorial covering agent parameters, execution controls, and scheduling options.

Overview​

Autonify agents can be executed in multiple ways:

  • Automatic Execution: Triggered after data source scans when changes are detected
  • Manual Execution: On-demand running via the Run Agent button
  • Scheduled Execution: Automated runs based on cron schedules
  • Programmatic Execution: Via API calls (for advanced integrations)

Viewing Active Agents​

Dashboard View​

The dashboard shows agent activity for your data source:

  • Navigate to your data source
  • Check the Agent Activity section
  • View which agents are active and operational

Agent Activity Dashboard

Sidebar Agent List​

The left sidebar displays all agents attached to your data source:

  • Each agent shows its name and description
  • Click on any agent to view details
  • Icons indicate agent type and status

Manual Agent Execution​

Running an Agent​

To manually execute an agent:

  1. Open the agent from the sidebar or agent list
  2. Click the Run Agent button
  3. The agent begins execution immediately
  4. Status updates to "Running..."

Run Agent Button

Execution Feedback​

After starting an agent:

  • Progress indicator shows the agent is running
  • Real-time status updates appear
  • Completion message confirms successful execution
  • Results are displayed once processing finishes

Canceling Execution​

While an agent is running:

  • The Run Agent button changes to a red Cancel button
  • Button shows Cancelling... while processing cancellation
  • Click Cancel to stop the execution
  • The agent will terminate gracefully
  • Partial results may be saved depending on the agent type

Viewing Execution History​

Run History​

Scroll down on any agent page to see:

  • Complete execution history
  • Timestamps for each run
  • Status indicators (Completed, Failed, Cancelled)
  • Duration of each execution

Execution Logs​

Each run includes detailed logs:

  • Processing steps performed
  • Tables or data analyzed
  • Any errors or warnings encountered
  • Performance metrics

Agent Run History

Managing Agents​

Accessing Agent Management​

  1. Click View All / Manage Agents at the bottom of the agent sidebar
  2. This opens the agent management interface

Note: Only Admin and Owner roles can see this option.

Agent Management Interface​

The management page has two tabs:

Connected Agents Tab​

  • Lists all agents currently attached to the data source
  • Each agent card shows:
    • Agent name and description
    • Category badge
    • View button to open agent details
    • Remove button to disconnect the agent
  • Removing an agent shows a confirmation dialog

Available Agents Tab​

  • Shows agents not yet connected to your data source
  • Each agent card displays:
    • Agent name and description
    • Category badge
    • Connect button to add the agent
  • Connected agents become immediately operational

Agent Management Interface

Agent Settings​

Accessing Agent Settings​

  1. Navigate to your data source
  2. Click on Agent Settings in the navigation
  3. Configure processing limits and exclusions

Processing Limits​

Control AI costs and processing time:

Table Processing Limit:

  • Slider to set maximum tables processed
  • Range: 1 to 100,000 tables
  • Shows current value next to slider
  • Description: "Maximum number of tables the AI will process before stopping"
  • Helps manage AI token consumption

Table Processing Limit

Exclusion Configuration​

Specify what to exclude from AI processing:

Selection Mode​

  • Include selected items only: Process only the selected tables
  • Exclude selected items: Process all tables except the selected ones

Table Selection Tabs​

Three tabs for different selection levels:

  1. Tables: Select specific tables individually
  2. Schemas: Select entire schemas to include/exclude
  3. Databases: Select complete databases to include/exclude

This configuration:

  • Improves processing speed
  • Reduces AI costs
  • Focuses agents on relevant data
  • Protects sensitive information

Exclusion Settings

Agent Scheduling​

Setting Up Schedules​

Configure automated agent execution:

  1. Navigate to your data source
  2. Open the scheduling interface for agents
  3. Click to create or edit a schedule for an agent
  4. The Schedule dialog opens

Schedule Options​

Schedule Type Selection​

  • Use Preset: Choose from predefined schedules
  • Custom Cron Expression: Enter custom cron syntax

Preset Schedules​

  • Every hour (at the start of every hour)
  • Every 6 hours
  • Every 15 minutes
  • Every 30 minutes
  • Daily at 2 AM
  • Daily at midnight
  • Weekly on Monday (at 2:00 AM)
  • Weekly on Sunday (at 2:00 AM)
  • Monthly on 1st (at 2:00 AM)

Custom Cron Expressions​

  • Create precise schedules using cron syntax
  • Example: 0 2 * * * (daily at 2 AM)
  • Supports complex patterns

Schedule Configuration Dialog​

  • Schedule Type: Dropdown to select preset or custom
  • Schedule Frequency: Dropdown with preset options (when using preset)
  • Cron Expression: Text input for custom cron (when using custom)
  • Description: Optional text area for schedule notes
  • Enabled: Switch to enable/disable the schedule
  • Save button to confirm changes

Agent Scheduling

Automatic Execution​

Trigger Conditions​

Agents run automatically when:

  • A data source scan completes
  • Changes are detected in the schema
  • New tables or columns are discovered
  • Data quality thresholds are crossed

Automatic Execution Behavior​

  • Agents run in a specific order based on dependencies
  • Failed agents don't block other agents
  • Notifications sent on completion or failure
  • Results available in agent history

Best Practices​

Execution Frequency​

  • Balance between data freshness and cost
  • Use scheduling for predictable workloads
  • Manual execution for ad-hoc analysis
  • Monitor execution history for optimization

Cost Management​

  • Set appropriate table processing limits
  • Exclude non-critical schemas/tables
  • Review agent logs to identify inefficiencies
  • Use scheduling during off-peak hours

Performance Optimization​

  • Exclude system tables and logs
  • Focus agents on business-critical data
  • Run heavy processing agents during low-usage periods
  • Monitor execution times and adjust limits

Troubleshooting​

Agent Won't Start​

  • Check if another execution is in progress
  • Verify agent is connected to data source
  • Ensure you have permission to run agents
  • Review any error messages in logs

Slow Execution​

  • Review table processing limits
  • Check exclusion settings
  • Consider scheduling during off-peak times
  • Examine logs for bottlenecks

Failed Executions​

  • Check execution logs for error details
  • Verify data source connectivity
  • Ensure sufficient permissions
  • Review agent-specific requirements

Next Steps​

After understanding agent execution:

  1. Configure agent parameters
  2. Learn about agent overview
  3. Manage data sources
  4. View agent results