Adrian Cole

Where are we going with observability?

In the last ten years, we've moved observability from concept to practice. We’ve focused on instrumentation and collection of signals, like metrics, logs, tracing and most recently profiling, as well capturing data in consistent ways. Now, we query our systems and pinpoint specific entities under investigation, such as slow requests through a Kubernetes service. Two years ago, ChatGPT forever changed how we interact with systems. We've moved from learning query languages to asking in plain english. We get context-specific advice and we can even ask the system to write queries for us! How, then, does AI impact our understanding or observability of systems? What type of work should we expect to have assistance with? Can AI help me figure out what I'm running, what happened in an outage, or even prevent one? This session is the now and future of AI native observability. You’ll learn what's possible through real examples of GenAI prompts. You'll understand why concepts like RAG work to your advantage with private information. You'll learn AI does all the hard work, even when using only cheap logs, in a real system. You’ll leave knowing how AI can help us decouple from experts and move more people into successful root cause analysis.