Where Are We Going With Observability?
summary
Argues that generative models practically augment observability by doing concrete work: use RAG with private docs to attach release notes and source code when parsing logs, run BERT-style NER to auto-redact PII, and use CaseGPT to translate CVE/debug strings into actionable Kubernetes steps. Shows Kibana integrations—Kibana AI assistant using an automatic import to extract fields from an S3 bucket—and cites LangTrace-style metrics: subjective success rate, token counts and cost (anecdotal CI runs at ~$20) with tokens as spend metric for billing and alerting.
AI-written summary · about
chapters (6)
- 0:00Observability and Gen AI
- 1:46Core observability signals
- 8:31Generative AI fundamentals
- 15:56Observing Gen AI systems
- 23:22AI tools for observability
- 31:08Future of AI in observability
related talks
28mWhere are we going with observability?
19mDo We Still Need to “Observe”? the Future of AI & Observability
29mElevating generative AI and LLM observability
32mGenerative AI for Observability: Automating Root Cause Analysis in Modern IT Systems
28mDO WE STILL NEED TO "OBSERVE" THE FUTURE OF AI & OBSERVABILITY
13mObservability Where are We Headed?
23mAI/ML + open source observability
31mTesting, Evaluating and Debugging Generative AI and Agentic Ap...