Leveraging AI and Automation in DevOps
summary
Claims that matter are concrete: instrument everything with OpenTelemetry and unify traces and metrics, run RAG pipelines and RAG with LLMs over RCAs and logs, fine tune self‑hosted LLMs (Llama3 mentioned) for private agents, and deploy an agent triage loop that iterates until a 99%‑confident fix before human approval. Demonstrations include automated ticket handling, cost‑optimization rules, and autoscaling suggestions; results reported are fewer manual RCAs, faster onboarding, and a roadmap toward autoscaling‑as‑a‑service.
AI-written summary · about
chapters (20)
- 0:00AI and DevOps introduction
- 2:31Infrastructure data for AI
- 6:03Data quality for AI
- 7:50AI value in DevOps
- 10:42AI for docs and knowledge
- 13:28AI for logs and RCA
- 15:57Fine-tuning AI models
- 17:03AI privacy and self-hosting
- 19:10Data pipelines for AI
- 20:56AI cost optimization
- 24:29AI and automation synergy
- 24:51Future of AI agents
- 28:28AI agent pricing
- 33:00AI ethics and legal issues
- 36:12AI accountability in postmortems
- 39:20Next AI in infrastructure
- 41:49Embedded AI systems
- 42:53AI handling incidents
- 43:34AI impact on junior roles
- 46:19Closing and Q&A
related talks
26mAI & SRE: Embracing the Ultimate Frontier of Automation
30mAutomate All the Things Panel Discussion
28mBeyond Dashboards: Architecting AI Agents for Autonomous Observability
15mIntelligent Incident Management in DevOps
50mHow to Test LLM Agents
29mLeveraging AI to Enhance DevEx & Operational Efficiency
25mDe-risking Deployments in an AI World
29mBuilding Trust in AI Agents