When OpenAI discovered they could reclaim 30,000 CPU cores simply by tuning the log-forwarding agent Fluent Bit—disabling a single function that ate ~35 % of one server’s cycles—something large and systemic became undeniable. In this episode of Pop Goes the Stack, #F5's Lori MacVittie, Joel Moses, and observability expert, Chris Hain, break down the hidden cost of telemetry in #AI-heavy architectures, why “logging is free” is a myth, and how modern systems demand a new breed of high-speed telemetry planes. Listen in to learn how Fluent Bit’s file-watching overhead compounded at scale, why profiling matters, and what enterprises can do now to control AI #observability costs. Chapters: 00:00 Welcome to Pop Goes the Stack 00:19 The most expensive lie: “logging is free” 02:25 Fluent Bit at scale: DaemonSet on every node (and the bill) 03:47 Inotify, memory churn, and the surprise CPU tax 04:43 Did turning off the feature actually solve the problem? 05:48 Finding it with profiling: System calls don’t lie 07:44 AI multiplies telemetry: More data, more destinations 09:02 Modern reality: Observability can cost more than the app 10:07 LLM logs aren’t “log lines” anymore (volume + metadata) 12:12 Reducing contention: Sampling, eBPF, buffering, offload 18:31 Enterprise key takeaways: Follow the money, drill down, optimize Read more about OpenAI's CPU recovery: https://go.f5.net/dvl5fk5d Learn how you can stay ahead of the curve and keep your stack whole with additional insights on app security, multicloud, AI, and emerging tech: https://go.f5.net/7cnnrgoi More about F5: https://go.f5.net/twngukhv Read our blog: https://go.f5.net/y9sdms2p Follow us on LinkedIn: https://go.f5.net/4itvvip6

 63      3