AI & SRE: Embracing the Ultimate Frontier of Automation
summary
Argues that LLM tool calling enables agentic workflows that call Splunk and other APIs to fetch logs, reason, and loop until a conclusion; proposes combining human-in-the-loop and AI-in-the-loop automation for safer remediation. Shows concrete builds: an alert triage agent that reads 500 errors and runs log queries, a request-routing flow that autofills ticket fields, and an incident similarity pipeline using embeddings + vector DBs; notes embedding creation cost is upfront while similarity queries stay cheap and fast.
AI-written summary · about
chapters (8)
- 0:00Introduction and background
- 3:09Audience AI usage and attitudes
- 5:15Evaluating AI like engineers
- 8:28Automation and human-in-the-loop
- 10:13AI in the loop automation
- 12:20Concrete AI use cases
- 20:07Summarization and burnout
- 21:55Q&A on embeddings and tool calling
related talks
46mLeveraging AI and Automation in DevOps
05mAI-Driven Incident Resolution – Hype or Reality
30mAI and Incident Management: How to Reduce Manual Work and Speed up the Incident...
26mBecoming AI engineers
29mLeveraging AI to Enhance DevEx & Operational Efficiency
13mDesigning AI-Intensive Applications
28mAI for Everything After Code
36mRisks and Opportunities of AI in Incident Management