From Firefighting to Zero-Touch
summary
Claims a reproducible path from reactive monitors to zero-touch AIOps by layering contextualized observability (logs/metrics/traces → topology/cause), multimodel AI (predictive + causal + generative), and automation. Demo uses Dino trace, Service now and Lipmos chaos to kill an account pod, scales replicas from 1 to 4, and restores SLOs after a 4–5 minute outage; business KPI (trade volume) drops are visible and normalized. Built workflow auto-creates incidents, runs a Flask remediation, and closes the incident when SLOs return.
AI-written summary · about
chapters (10)
- 0:00Introduction and Agenda Overview
- 1:17Key Challenges in SRE and Modern IT
- 3:04The Pitfalls of Automation Without Observability
- 4:09The Importance of Contextualized Observability
- 6:28Integrating AI and Automation for Predictive Maintenance
- 7:11Connecting Business and Operations for End-to-End Monitoring
- 8:29Accenture's Journey to Zero Touch AIOps
- 11:52Demo: Implementing Zero Touch AIOps
- 13:26Live Demo: Chaos Experiment and Self-Remediation
- 22:33Conclusion and Lessons Learned
related talks
27mAI-Powered Reliability: Stop Fighting Fires, Prevent Them
46mObservability in the real world: Improve systems w/iterative approach
14mFrom Dashboard to Defense: Automating Resilience at
41mCustomer Centric Observability at Scale Leveraging AIOps & OpenTelemetry
25mAI-Powered O11y in Microservices Architectures
29mObservability; a gentle introduction
18mChaos Experiments under the lens of AIOps
19mFrom Alerts to Action: AI-Powered Incident Response Systems