The Everything Feed - All Packet Pushers Pods Tech Bytes: Monitoring AI Agents and MCP Servers with Cisco ThousandEyes (Sponsored)
Nov 17, 2025
Joe Doherty, a Product Solutions Architect at Cisco ThousandEyes, dives into the fascinating world of AI agents and Model Context Protocol (MCP) servers. He explains how ThousandEyes provides crucial visibility into network performance for AI models and services. Topics include agentic AI use cases like automated refunds, key performance metrics to monitor, and the role of DNS and network dependencies. Joe emphasizes that familiar network skills still apply, making it accessible for engineers venturing into AI monitoring.
AI Snips
Chapters
Transcript
Episode notes
End-to-End Visibility Is Core
- Cisco ThousandEyes provides end-to-end visibility from users to applications across enterprise, internet, and cloud networks.
- That visibility helps identify, remediate, and optimize user experience for both owned and third-party apps.
Agents Combine LLMs And Tool Calls
- Agentic AI pairs LLM inference with tool calls, often via the emerging MCP standard.
- MCP provides context and tools so agents can act on internal and third-party data programmatically.
Food Delivery Discount Example
- Joe contrasts old rule-based chatbots with modern agents that access data and perform actions, like auto-applying discounts during a delivery disruption.
- That example shows agents can act on context rather than just pointing users to forms.