Azure Updates: Resilience; Microsoft Foundry; AI agents

Microsoft CTO Mark Russinovich teamed up with Adam Bogobowicz and Molina Sharma to discuss architecting for resilience. Despite capabilities like availability zones, health probes often turn up issues like a single dependency resulting in a point of failure. Traditional disaster recovery planning has usually focused on infrastructure as the most important dependency. Today, it is more commonly AI models or inference endpoints. 

“There is a second shift underneath the first. An architecture diagram assumes a human drew it and a human will read it. Both of those assumptions are ending, and the dependencies it describes are no longer all deterministic. That changes what it means to know your estate is resilient,” Russinovich wrote. He encouraged customers to “start resilient,” focusing on best practices laid out in resources like the Cloud Adoption Framework. 

Microsoft partner group product manager Naomi Moneypenny looked at how to deploy production agents in Microsoft Foundry. The company is expanding its GPT series of models with the addition of GPT-6 Sol and GPT-6 Luna as part of the generally available model line up in Foundry. Closely related capabilities, like newly available GPT-6 Astra introduce advanced reasoning for software engineering. GPT-6 Sol and its predecessor model GPT-5.6 Sol offer cost-effective capabilities for enterprise agents, whereas GPT-6 Luna is more targeted toward high-volume work.