Empowering SREs to Predict and Prevent Downtime
summary
A deployed adaptive reasoning platform called Thorus integrates with existing observability tools, ingests metrics logs and traces, and autonomously maps service dependencies to simulate blast radius before deploys. It claims to diagnose issues up to 90% faster, cut observability waste where ~98% of ingested data is irrelevant, and run lightweight in‑house models that are over 80% accuracy in customer tests while recommending ordered mitigation steps and reducing MTTR and cost compared with routing to large LLMs.
AI-written summary · about
chapters (7)
- 0:00Introduction and Speaker Background
- 1:27The Birth of Thoris AI
- 2:21Challenges in Achieving High Availability
- 5:23Adaptive Reasoning Technology
- 9:22Use Case: Sally's On-Call Nightmare
- 13:21Benefits of Adaptive Reasoning
- 15:54Q&A Session
related talks
36mSRE'ing the SRE AI Agents
40mWhen Incidents Fix Themselves: AI SRE in action
12mSRE Principles for High-Speed Infrastructure Monitoring
21mMaximizing resilience through SRE and governance: A holistic approach
14mFrom Dashboard to Defense: Automating Resilience at
32mSTOP TREATING AI SRE AS JUST AN ENGINEERING PROBLEM
46mObservability in the real world: Improve systems w/iterative approach
54mUnraveling the Differences