BRK452 Breakout Wed 18 Nov 13:45 PST
AI-powered resiliency for modern workloads with Mark Russinovich
Mark Russinovich, Amit Ganguli, Surya Sripathi Raju, Enrique González Bejarano
Summary
From the recording / Microsoft AI summary
Azure Infrastructure Resiliency Manager was announced as generally available at Ignite, and backup to a region of your choice is now live. The session said Azure will not move workloads to another zone or region automatically during an outage, so resilience has to be designed in. Several disaster recovery and cyber recovery features were described as already in public preview, without naming which.
Session description
Microsoft's catalogue description
Join Mark Russinovich and Azure resiliency experts to explore how organizations can engineer for failure as they modernize and adopt AI. Learn how customers use Azure to identify critical dependencies, assess resilience gaps, strengthen recovery capabilities, and modernize infrastructure. Attendees will gain practical strategies for building resilient architectures, improving business continuity, and preparing applications and data to better withstand disruption.
Demo timestamps
Microsoft's player does not support links to a specific time, so open the recording and skip to the time shown.
- Skip to 0:20:08
Resiliency Agent with an uploaded architecture diagram
From the portal Resiliency Overview, Azure Copilot's Resiliency Agent is given an architecture diagram and asked to configure zonal resiliency. It identifies the resources and asks follow-up questions. Infrastructure-as-code generation and Service Groups are also covered.
- Skip to 0:25:28
Portal resiliency experience at scale
Covers the landing overview, infrastructure resiliency, and Service Group tabs (criticality, goals, drills, top recommendations). Also shows an at-scale view of zonal posture, including when service groups have not been created.
- Skip to 0:27:00
Cost, downtime and risk alongside zonal goals
Criticality tier and zonal goals personalise recommendations. The agent is asked for cost and downtime implications by resource type, with a deeper report and risk-assessment view. Regional resiliency (RPO/RTO) is described as coming shortly after Ignite.
- Skip to 0:30:40
Backup status and unprotected VMs
Focus moves to data resiliency and cyber recovery. The at-scale view switches to backup status, unprotected VMs are listed, and Copilot can be asked to help configure backup.