< session />
From Noise to Signal: Using MCP Servers for AI-Driven Alerting and Monitoring
Traditional monitoring and alerting systems generate too much noise and too little insight. They detect issues but often miss the context that matters most. The Model Context Protocol (MCP) changes this by bringing structure, context, and reasoning to observability workflows. In this session, you will see how MCP servers can transform alerting and incident management into proactive, intelligent systems.
Through real-world examples and live demonstrations, the session shows how MCP can feed contextual data, such as alert metadata, recent deployments, logs, and runbooks, into AI models to create responsive agents that explain what went wrong, why it happened, and how to fix it. Attendees will also see an AI-powered Alert Assistant in action, learn the underlying architecture, and take away templates ready for production adaptation.
What You Will Learn
-
What Model Context Protocol (MCP) is and how it differs from prompt engineering
-
How to integrate AI agents into observability pipelines for context-aware alerting and triage
-
Practical use cases including alert correlation, root cause analysis, and automated post-mortems
-
How to build and deploy an AI-driven Alert Assistant with reusable templates and manifests
Who Should Attend
Developers, SREs, DevOps engineers, and software architects interested in applying AI and MCP to improve observability, reduce noise, and automate incident management.
< speaker_info />
About the speaker
Devyani Kota
Senior Software Engineer, Red Hat
Devyani is a Senior Software Engineer at Red Hat, specializing in distributed systems and cloud-native applications. With over 8 years of experience in the software industry, she has amassed a wealth of expertise across multiple sectors, fintech, site-reliability, and data engineering. In addition to that, she has a passion for open source software where she contributed to Fedora, and GlusterFS projects in the past and has shared her knowledge at multiple local meetups.
