​

Dashboard Overview

​

Why Monitor Your Agents?

• Track Performance: Understand how your agents are performing across all conversations
• Identify Issues: Catch errors, failures, and quality problems before they impact users
• Optimize Costs: Monitor usage patterns to optimize LLM, voice, and infrastructure spending
• Improve Quality: Analyze conversation patterns to refine prompts and configuration
• Scale Confidently: Track concurrency and queue metrics to plan capacity growth
• Respond Quickly: Detect and resolve issues in real-time with live metrics

​

Accessing the Dashboard

1. Direct Access: Click the Dasha BlackBox logo from any page
2. Main Menu: Select “Dashboard” from the top navigation
3. URL: Navigate to https://blackbox.dasha.ai/ (or your deployment URL)

​

Dashboard Sections

​

1. Welcome Header with Timeframe Selector

• Personalized greeting: “Welcome back!”
• Descriptive status: “Here’s what’s happening with your AI agents today”
• Timeframe Selector: Filter all metrics by time period
  • 24 hours: Recent activity and real-time trends
  • 7 days: Weekly patterns and performance
  • 30 days: Monthly overview and historical trends

• Affects all dashboard metrics simultaneously
• Updates call counts, statistics, and recent activity
• Calculates rolling windows (e.g., last 24 hours from now)
• Persists across page refreshes

​

2. Statistics Grid

​

3. Agent List

​

Agent Card Components

• Agent Name: Descriptive name in large, bold text
• Status Badge: Green “Enabled” or gray “Disabled”
• Quick Actions Menu (three-dot icon):
  • View Details: Open agent detail page
  • Edit Agent: Navigate to configuration editor
  • View Calls: Filter calls page by this agent
  • Delete Agent: Remove agent (requires confirmation)
• Link Icon: Direct access to agent detail page

• System Prompt Summary: First 2 lines of your agent’s system prompt
• Created Date: When the agent was first created
• Last Updated: Most recent configuration change

• Total Calls: All calls for this agent in selected timeframe
• Active Calls: Currently running conversations
• Completed Calls: Successfully finished conversations

​

Empty State

• “No agents found” message
• “Create Your First Agent” button
• Links directly to agent creation wizard

​

Quick Action Cards

• Indigo background with plus icon
• “Set up in 30 seconds” subtitle
• Creates new agent draft and opens editor

• Emerald background with link icon
• “Manage your agents” subtitle
• Links to full agent management page

​

4. Recent Calls Sidebar

• Status Indicator: Colored dot (blue=running, green=completed, yellow=pending, red=failed)
• Call ID: Truncated identifier (first 12 characters)
• Agent ID: Which agent handled the call
• Endpoint: Phone number or “web” for web calls
• Timestamp: When the call completed (or was created)
• Status Badge: Color-coded status label

• Entire call row is clickable
• Opens Dasha Inspector in new tab
• External link icon appears on hover
• Deep dive into conversation details

​

5. Call Status Summary

1. Running (blue dot): Active conversations in progress
2. Pending (yellow dot): Scheduled, waiting for available slot
3. Queued (orange dot): Waiting for concurrency capacity
4. Completed (green dot): Successfully finished calls
5. Failed (red dot): Calls that encountered errors
6. Canceled (gray dot): Manually canceled or deadline-expired calls
7. Created (purple dot): Just created, not yet processed
8. Unknown (gray dot): Status not recognized (rare)

• Identify which statuses dominate your traffic
• Spot abnormal failure rates
• Monitor queue buildup
• Validate call lifecycle progression

​

6. Concurrency Alert Banner

• Approaching concurrency limit (above 70% utilization)
• At concurrency limit (100% utilization)

• Warning icon with alert message
• “Contact support” link for tier upgrades
• Non-intrusive banner style (not blocking)

​

Real-Time Metrics

​

What Updates in Real-Time

• Reflects current active call count
• Updates when you click the refresh button
• Manual refresh only (no automatic polling)

• Enabled/disabled state reflects current configuration
• Updates when you enable or disable agents

​

What Updates on Timeframe Change

• Total calls recalculated for new date range
• Pending calls updated for new window
• Recent calls list refreshed with new timeframe

• Per-agent call counts recalculated
• Completed/active breakdown updated

​

Manual Refresh Points

• Click refresh icon in tooltip to update immediately
• No automatic polling to avoid unnecessary API calls

• Refresh by changing timeframe
• Reloading the page fetches latest data

​

Filtering and Search

​

Timeframe Filtering

• 24 Hours: Recent activity, real-time troubleshooting
• 7 Days: Weekly trends, pattern identification
• 30 Days: Monthly overview, historical analysis

• All metric cards (Total Calls, Pending Calls)
• Recent calls list
• Call status summary
• Per-agent statistics

1. Click desired timeframe button (24h, 7d, 30d) in header
2. All metrics update automatically
3. Selection persists across page refreshes

​

Advanced Filtering

• Click “View Calls” in agent card menu
• Filters calls page to show only that agent’s calls

• Navigate to Calls page for advanced search
• Search by call ID, endpoint, status
• See Call Management for details

• Currently limited to preset ranges (24h, 7d, 30d)
• Custom date ranges available via API

​

Dashboard Widgets

​

Widget Overview

​

Customization

​

Current Customization Options

• Choose 24h, 7d, or 30d views
• Persists across sessions

• Dark mode (default and current only option)
• Light mode prepared for future release

​

Future Customization (Planned)

• Drag-and-drop widget reordering
• Hide/show specific widgets
• Resize widget panels

• Select specific agents to track
• Create custom date ranges
• Add custom metric cards

• Set custom threshold alerts
• Email/webhook notifications for metrics
• Configurable alert conditions

​

Best Practices

​

Daily Monitoring

• Concurrency utilization under 70%
• No unusual spike in failed calls
• Pending calls queue is manageable (under 10% of capacity)
• All critical agents are enabled

• Review recent calls for errors (red status dots)
• Check agent statistics for unusual patterns
• Verify total call count aligns with expectations

​

Weekly Monitoring

• Compare 7-day metrics to previous week
• Identify agents with declining usage
• Review failure rate trends (should be under 5%)
• Check concurrency peaks during business hours

• Identify unused or disabled agents (candidates for deletion)
• Review agent creation dates (clean up old test agents)
• Check for agents with zero calls in 7 days

• Review peak concurrent call times
• Plan concurrency tier upgrades if approaching limits
• Estimate monthly usage for cost planning

​

Monthly Monitoring

• Review 30-day total calls and growth trends
• Analyze completed vs failed call ratios
• Compare agent performance across entire portfolio
• Identify cost optimization opportunities

• Review all agents for outdated configurations
• Clean up test or unused agents
• Consolidate duplicate agents if possible
• Update system prompts based on conversation learnings

• Review LLM usage patterns (switch to cheaper models if possible)
• Analyze voice provider costs vs quality
• Consider prompt optimization to reduce token usage
• See Best Practices for optimization tips

• Forecast next month’s capacity needs
• Plan new agent deployments
• Review webhook and integration health
• Update monitoring and alerting strategies

​

Proactive Alerting

• Concurrency utilization exceeds 80%
• Error rate exceeds 10% in any 1-hour period
• Specific critical agents go offline
• Webhook endpoints become unreachable

// Monitor FailedWebHookPayload events via webhook
app.post('/webhooks/blackbox', (req, res) => {
  const payload = req.body;

  if (payload.type === 'FailedWebHookPayload') {
    failureCount++;

    // Alert if 10+ failures in 1 hour
    if (failureCount > 10) {
      sendAlert('High failure rate detected');
    }
  }

  res.status(200).send('OK');
});

​

Dashboard Hygiene

• Delete test agents after validation
• Archive or disable unused agents
• Use descriptive agent names (“Customer Support - English” not “Agent 1”)
• Maintain under 20 agents for dashboard performance

• Completed calls are retained for 30 days by default
• Use Calls page to review and analyze historical data
• Export important transcripts before they expire

• Set calendar reminders for weekly and monthly reviews
• Document findings in a monitoring log
• Share insights with your team via Slack or email

​

Troubleshooting Common Issues

​

Metrics Not Updating

1. Refresh the page (hard reload: Cmd+Shift+R or Ctrl+Shift+F5)
2. Try changing timeframe to trigger data refetch
3. Clear browser cache and cookies
4. Check browser console for API errors

• Cached data from previous session
• API timeout or rate limiting
• Network connectivity issues

​

Agent Statistics Show Zero

1. Change timeframe to 7d or 30d to expand date range
2. Check if calls are for a different agent
3. Verify agent ID matches in Calls page
4. Wait 1-2 minutes for statistics to propagate

• Calls outside selected timeframe
• Statistics calculation delay (rare)
• Agent was recently created

​

Concurrency Indicator Not Refreshing

1. Click refresh icon in concurrency card tooltip
2. Reload the page
3. Check if concurrency feature is enabled for your account

• Manual refresh required (no auto-polling)
• API connectivity issue
• Concurrency monitoring disabled

​

Recent Calls Not Appearing

1. Expand timeframe to 7d or 30d
2. Check Calls page for more detailed view
3. Verify calls have completed (calls in progress may not appear)

• All calls outside selected timeframe
• Calls still running (appear after completion)
• Search/filter applied on Calls page

​

Next Steps

​

Essential Monitoring Pages

• Call Management - View scheduled and past calls
• Call History - Search and filter all historical calls
• Call Transcripts - Read conversation transcripts

• Call Inspector Deep Dive - Debug call issues
• Analytics - Advanced performance metrics
• Agent Performance - Per-agent optimization

• Concurrency Monitoring - Track concurrent call limits

​

Related Configuration

• Webhooks Overview - Set up event notifications
• Webhook Events - Available event types
• Webhook Security - Secure your webhooks

• Best Practices - Improve agent quality and reduce costs
• Advanced Features - Enable post-call analysis and logging

​

API Cross-Refs

• GET /api/v1/agents - List agents
• GET /api/v1/calls/statistics/statuses - Call statistics
• POST /api/v1/callresults/search - Search call results
• GET /api/v1/misc/concurrency - Concurrency metrics

• GET /api/v1/agents/ - Get agent details
• PUT /api/v1/agents/ - Update agent
• DELETE /api/v1/agents/ - Delete agent

• GET /api/v1/calls/queue/list - List queued calls
• DELETE /api/v1/calls/ - Cancel call