Agent Framework

<h2 id="agent-framework---agent-overview">Agent Framework - Agent Overview</h2> <p><img class="lazyload d-inline-block" data-src="https://github.com/1w2w3y/grafana-dashboards/raw/master/agent-framework/agent-framework-grafana.gif" alt="Agent Framework Dashboard"/></p> <p>Comprehensive monitoring dashboard for <strong>Microsoft Agent Framework</strong> performance, tracking operations, token usage, costs, errors, and detailed trace analysis. Monitor your AI agents&rsquo; health, response times, success rates, and resource consumption in real-time.</p> <h2 id="-what-is-microsoft-agent-framework">📊 What is Microsoft Agent Framework?</h2> <p>The <a href="https://github.com/microsoft/agent-framework" target="_blank" rel="noopener noreferrer">Microsoft Agent Framework</a> is an open-source development kit for building AI agents and multi-agent workflows for .NET and Python. It brings together the best of <a href="https://github.com/microsoft/semantic-kernel" target="_blank" rel="noopener noreferrer">Semantic Kernel</a> and <a href="https://github.com/microsoft/autogen" target="_blank" rel="noopener noreferrer">AutoGen</a>, combining their strengths while adding new capabilities for enterprise-grade AI applications.</p> <h2 id="-why-this-dashboard">🎯 Why This Dashboard?</h2> <p>As you build AI agents that make autonomous decisions, call external tools, and process user requests, you need visibility into:</p> <ul> <li><strong>Performance</strong>: Are your agents responding quickly enough?</li> <li><strong>Reliability</strong>: What&rsquo;s your success rate? Where are errors occurring?</li> <li><strong>Cost</strong>: How much are your AI operations costing in token consumption?</li> <li><strong>Usage</strong>: Which agents are most active? When are peak usage times?</li> <li><strong>Debugging</strong>: What&rsquo;s happening inside failed operations?</li> </ul> <p>This dashboard answers all these questions and more—in real-time.</p> <h2 id="-dashboard-highlights">✨ Dashboard Highlights</h2> <h3 id="-performance-monitoring">📈 Performance Monitoring</h3> <ul> <li><strong>Response Time Trends</strong>: Track average, 95th, and 99th percentile latencies</li> <li><strong>Success Rate Gauge</strong>: Instant health check with color-coded thresholds</li> <li><strong>Throughput Analysis</strong>: Monitor operations per minute to identify peak loads</li> <li><strong>Agent-Level Metrics</strong>: Compare performance across different agents</li> </ul> <h3 id="-token-usage--cost-analysis">💰 Token Usage &amp; Cost Analysis</h3> <ul> <li><strong>Token Consumption Over Time</strong>: Stacked visualization of input/output tokens</li> <li><strong>Cost Estimation</strong>: Daily spending estimates based on model pricing</li> <li><strong>Per-Agent Breakdown</strong>: Identify which agents consume the most resources</li> <li><strong>Trend Analysis</strong>: Spot usage patterns and optimize accordingly</li> </ul> <h3 id="-error-analysis--debugging">🔍 Error Analysis &amp; Debugging</h3> <ul> <li><strong>Error Rate Monitoring</strong>: Overall system health at a glance</li> <li><strong>Error Timeline</strong>: Identify when problems occur</li> <li><strong>Recent Errors Table</strong>: Quick access to the latest failures</li> <li><strong>Root Cause Analysis</strong>: Drill down with agent name, error type, and duration</li> </ul> <h3 id="-detailed-trace-analysis">🔬 Detailed Trace Analysis</h3> <ul> <li><strong>Interactive Trace Selection</strong>: Click any operation to see full execution details</li> <li><strong>Trace Visualization</strong>: Complete execution flow with timing and dependencies</li> <li><strong>Span Analysis</strong>: Understand the hierarchical relationships in your agent operations</li> <li><strong>Performance Deep-Dive</strong>: Identify bottlenecks in complex agent workflows</li> </ul> <h3 id="-agent-activity-insights">📊 Agent Activity Insights</h3> <ul> <li><strong>Agent Performance Summary</strong>: Comprehensive table with operations, duration, and success rates</li> <li><strong>Utilization Heatmap</strong>: Visual representation of agent activity over time</li> <li><strong>Response Time Percentiles</strong>: Compare latency distribution across agents</li> </ul> <h2 id="-customization">🔧 Customization</h2> <h3 id="adjust-cost-estimation">Adjust Cost Estimation</h3> <p>The dashboard uses example pricing. Update the cost calculation query to match your model pricing:</p> <div class="code-snippet "><div class="lang-toolbar"> <span class="lang-toolbar__item lang-toolbar__item-active">kql</span> <span class="code-clipboard"> <button x-data="app_code_snippet()" x-init="init()" @click="copy()"> <img class="code-clipboard__icon" src="/media/images/icons/icon-copy-small-2.svg" alt="Copy code to clipboard" width="14" height="13"> <span>Copy</span> </button> </span> <div class="lang-toolbar__border"></div> </div><div class="code-snippet "> <pre data-expanded="false"><code class="language-kql">// In the &#34;Daily Cost Estimation&#34; panel query input_cost = input_tokens * 0.00000025, // Your input token rate output_cost = output_tokens * 0.000002 // Your output token rate</code></pre> </div> </div> <h2 id="-contributing">🤝 Contributing</h2> <p>Found a bug or have a feature request?</p> <ul> <li><a href="https://github.com/1w2w3y/grafana-dashboards/issues" target="_blank" rel="noopener noreferrer">Open an issue</a></li> <li>Submit a pull request</li> <li>Share your customizations!</li> </ul> <h2 id="-resources">📚 Resources</h2> <ul> <li><a href="https://learn.microsoft.com/en-us/agent-framework/overview/agent-framework-overview" target="_blank" rel="noopener noreferrer">Microsoft Agent Framework Documentation</a></li> <li><a href="https://github.com/microsoft/agent-framework" target="_blank" rel="noopener noreferrer">Agent Framework GitHub Repository</a></li> <li><a href="https://learn.microsoft.com/en-us/azure/azure-monitor/app/app-insights-overview" target="_blank" rel="noopener noreferrer">Azure Application Insights</a></li> <li><a href="/docs/">Grafana Documentation</a></li> </ul> <hr /> <p><strong>Ready to gain visibility into your AI agents?</strong> Import this dashboard and start monitoring your Agent Framework applications today! 🚀</p>