Production Monitoring with Codex: Grafana, Kubernetes, & Security
- Source
- youtube.com
- Author
- OpenAI
- Date
Context
OpenAI published a video of three production-monitoring demos using Codex, its . Hosts Tony Loehr and Anke Hao, per the description, start from a 3 a.m. scenario of climbing checkout errors and follow the failure in Grafana, a dashboard for system metrics.
The second demo investigates a rollout on Kubernetes (software that runs applications in containers) whose containers restart because they run out of memory. The third links a finding from Codex Security to service availability through a missing resource limit, a setting that caps how much of a resource, such as memory, a container may use. OpenAI says the demos show how telemetry, release history and code support an investigation, and how engineers review a proposed change and check the result.
What this establishes is the workflow OpenAI chose to demonstrate: an assisting while engineers review. The description gives no data on how often Codex finds the right cause.
- Agent mode — A mode in which one agent can inspect, edit, run commands, and iterate toward an outcome.
- Tool — An operation that the software around a model makes available for the model to request, with a name, expected inputs and a kind of result.
- Agent — A system in which a model repeatedly chooses operations, a harness carries them out, and the results return to the model to inform its next choice.
Checking sign-in…
Loading comments…






