OpenAI: Production Monitoring with Codex: Grafana, Kubernetes, & Security
In this technical demonstration from OpenAI, cohosts Tony Loehr and Anke Hao present three practical production-monitoring scenarios utilizing Codex. The walkthrough begins with diagnosing an escalating checkout failure using Grafana telemetry. They then investigate an unstable Kubernetes rollout responsible for causing out-of-memory restarts. Additionally, the presentation explores how a Codex Security finding connects overall service availability to an unconfigured resource limit. Throughout these examples, the hosts illustrate how telemetry, release history, and source code support engineering investigations. Finally, they demonstrate the workflow engineers use to review proposed remediation changes and verify the operational outcome.
Original sources
YouTube OpenAI
Video:Production Monitoring with Codex: Grafana, Kubernetes, & Security ↗
AI Radar wrote this summary from the publisher’s notes with AI, not from a transcript. How it works.