A Deep Dive into FinOps Integration with Kubernetes for Enhanced
summary
Claims an end-to-end FinOps pipeline that collects provider APIs, normalizes into the FinOps Foundation FOCUS conversion via LLMs, and stores records in a central data lake with a demonstrated Delta table upload so the ETL actually runs. Uses Kubernetes CRDs and operators to spawn operator-driven exporters and Prometheus pipeline, runs a schema‑mapping experiment across AWS GCP Oracle where GPT‑4 gives best results, and feeds analyses back to autoscalers to enact changes automatically.
AI-written summary · about
chapters (13)
- 0:00FinOps and Kubernetes context
- 5:20What is FinOps
- 7:06FinOps cost and usage
- 9:18FinOps integration challenges
- 10:44Automated FinOps framework
- 15:00Data collection with operators
- 17:49Standardizing cost data with LLMs
- 23:53Prometheus cost data storage
- 25:00Cost optimization with metrics
- 27:12LLMs for data analysis
- 28:57Live demo of framework
- 30:23Conclusion and future
- 31:25Q&A
related talks
46mFinOps: Optimize for K8s
38mA story about platform engineering — Niklas Voss
36mDataOps and MLOps: NLP Development and Deployment in Kubernetes made easy
53mImplementing Kubernetes on with confidence
22mBuilding a MLOps Platform with Kubeflow on GKE
35mWhy Kubernetes Can't Get Around FinOps – Cost Management Best Pra...
33mApache Kafka Devops with Kubernetes and Gitops
39mOperating Deep Learning Pipelines Anywhere Using Kubeflow