AI-Driven Observability and Operations in Cloud-Edge Systems
summary
An experimental AIOps framework ties OpenNebula+Prometheus+Grafana to machine‑learning models that produce per‑VM hourly CPU forecasts and feed placement and elasticity decision modules. The prototype generates placement suggestions only (no automatic migration), reports a high accuracy percentage from lab traces, and reduces suggested migrations and contention in on‑prem clusters. Next steps are automated enforcement, memory and network predictions, anomaly detection, and packaging the framework with the OpenNebula distribution.
AI-written summary · about
chapters (14)
- 0:00AI-driven observability intro
- 1:34What is observability
- 3:03AI's role in observability
- 4:30Open source vs vendor lock-in
- 5:54AI Ops framework components
- 7:02OpenNebula hybrid cloud
- 9:34Prometheus monitoring
- 11:00AI workload prediction
- 13:28AI Ops architecture
- 16:21Prototype with Grafana
- 17:25Future AI Ops plans
- 18:30Contribute to project
- 18:53EU-funded project
- 19:35Closing & Q&A
related talks
26mAI-Driven Cloud Observability: Reduce toil and amplify time to resolution
42mAIOps for Distributed Environments - Deep Dive
20mGenAIOps - OpenSearch for AI & GenAI Platform Observability
37mAIOps: Anomaly Detection with Prometheus and Istio
30mIntelligent Analysis on Cloud Infrastructure With AIOps
19mAIOps
22mAI Observability At Meta Scale
24mDemo of Kubernetes at the Edge for AI Model Deployments