Oh, Bring Back My Logging Log Anomaly Detection with Opni
summary
Opni is an open‑source observability stack that applies pre-trained models to Kubernetes control plane, Rancher and Longhorn logs, using a Gateway + agent architecture with an OpenSearch backend, Cortex/Grafana for metrics and Alertmanager; it parses logs with the Drain parser and kicks off a NeuLog seq2seq deep model (GPU required) only after steady state. A k3s outage demo shows Opni flagging anomalies at 10:29 with the incident logged at 10:31, and model training taking ~30 of 45 minutes.
AI-written summary · about
chapters (9)
- 0:00Introduction and background
- 2:22Agenda and project overview
- 3:59The observability landscape
- 7:28Opni capabilities and AIops
- 11:24Technical architecture
- 14:04Deployment and setup
- 16:19AI methods and use cases
- 21:12Installation and demo
- 23:35Roadmap and conclusion
related talks
31mAnomaly Detection on OpenStack Logs using M.L.
43mBeyond Logs: Real-Time Anomaly Detection with Machine Learning
12mOpen-source observability is the way forward, and here’s why
39mDemystifying Anomaly Detection in OpenSearch
22mFrom Telemetry to Insight: Building AI-Powered Observability Pipelines with OpenSearch
20mGenAIOps - OpenSearch for AI & GenAI Platform Observability
21mKubernetes Under the Magnifying Glass: Observability With OpenSearch
27mAlert Fatigue to Action: Observability as the Backbone of Practical AIOps with OpenSearch