Accelerating Incident Response With OpenTelemetry and MCP: Building...
summary
Shows an architecture where OpenTelemetry collectors feed logs, metrics and traces into an MCP server that runs Grafana, Loki, Prometheus and Tempo, uses LLMs (ChatGPT/perplexity or custom models) to detect anomalies, and automatically creates incidents, Slack channels and dashboards; demo builds a dashboard and posts an incident from an AKS cluster in ~10 minutes. Claims 20–50% MTTR reduction, supports Grafana MCP server and MCP super assistant, and describes using a vector DB for RAG over past Jira tickets to auto-suggest fixes.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 1:44OpenTelemetry for observability
- 3:50Incident response challenges
- 5:57Model Context Protocol benefits
- 9:28Grafana MCP server demo
- 10:12Live Grafana MCP demo
- 22:34Automating incident response
- 24:17Vector search for incidents
- 26:49Production adoption outlook
- 27:00Q&A
related talks
26mExploring Observability with MCP Servers
36mModernizing Incident Response with LLMs, RAG, and the MCP
17mOpenTelemetry Looks Good To Me: Loki, Grafana, Tempo, Mimir
25mLet Me Be Your OpenTelemetry Champ
19mBest Practices for Rapid Insights from Metrics, Logs, Traces and Profiles
18mBuilding an APM with OpenTelemetry and OpenSource
21mSpeed Up Troubleshooting with Integrated Application and Infrastructure Observability
48mDevOps Transformation: Introducing Incident Management and Maximizing Monitoring Value