Skip to main content

Monitoring Dashboard

Monitor your Bindu agents with real-time dashboards that provide insights into performance, errors, and usage patterns.

Dashboard Options

Grafana

Create custom dashboards with:
  • Task Metrics - Throughput, latency, error rates
  • Agent Performance - Execution time, success rates
  • System Health - Active tasks, memory usage
  • Alerts - Automated notifications for issues

Jaeger UI

Built-in visualization for:
  • Trace Timeline - Visual span hierarchy
  • Service Dependencies - System architecture
  • Performance Stats - Latency percentiles
  • Error Analysis - Exception tracking

Custom Dashboards

Build your own with:
  • Prometheus + Grafana
  • Datadog
  • New Relic
  • Honeycomb

Key Metrics to Monitor

Performance

  • P50/P95/P99 Latency - Response time percentiles
  • Throughput - Requests per second
  • Error Rate - Percentage of failed requests

Health

  • Active Tasks - Currently processing
  • Queue Depth - Pending tasks
  • Memory Usage - Resource consumption

Business

  • Agent Usage - Most/least used agents
  • User Activity - Request patterns
  • Cost - LLM API usage and costs

Alerting

Set up alerts for:
  • High error rates (> 5%)
  • Slow response times (> 5s)
  • Queue buildup (> 100 tasks)
  • Memory pressure (> 80%)

Next Steps