Agent Performance Monitor

You get real-time metrics collection across system resources, agent performance, and task execution, plus bottleneck detection that correlates issues across CPU, memory, I/O, and coordination layers. The SLA monitoring is comprehensive with configurable thresholds and automated alerting when availability, response time, or error rates drift out of bounds. Pattern recognition tracks recurring bottlenecks so you can catch systematic issues before they cascade. Most useful when you're running multi-agent swarms and need visibility into where coordination breaks down or resources get starved. The metrics collection is genuinely multi-dimensional, tracking everything from core utilization to inter-agent communication overhead.