Operations | Monitoring | ITSM | DevOps | Cloud

Azure Virtual Desktop (AVD) is rapidly growing in popularity as modern way to deliver virtual desktops and apps to users, with Azure providing the infrastructure as alternative to on-prem VDI environments. As organizations expand their use of AVD in Azure, monitoring becomes critical.

When engineering teams evaluate AI agents, the first questions are usually about capability. Which model performs best? How much faster can it write code? What's the return on investment? Security, if it enters the conversation at all, tends to come later. Patrick Dawkins, Principal Software Engineer at Upsun, thinks that's backward. Over the past year, he's been building the infrastructure that enables AI agents to operate safely within engineering teams.

In 1846, Britain poured roughly 7% of its national income into railways, proportionally about three times what the U.S. spends on AI infrastructure today. The technology delivered everything it promised, and a generation of investors still lost their shirts. What sorted the winners from the wreckage wasn't conviction about the technology; it was whether ROI was measured or asserted. The man who fixed that problem gave his name to the S&P 500.

Hundreds of files vanished from SharePoint. How to tell ransomware or a wipe from a user re-filing a folder, using the signals in the audit log that actually separate them.

AI cost reduction means lowering what you spend to run AI (tokens, inference, and compute) without sacrificing quality. The highest-leverage tactics, prompt caching, batching, and routing easy work to smaller models, cut spend 50 to 90% by removing waste, not capability. Somewhere right now, a finance leader is opening an AI bill that has quietly tripled, with no new product to show for it. Nobody approved it. No single decision caused it.

If your team builds with AI, someone is defaulting to the biggest model available (say, Fable) because it feels like the safe pick, and the safe pick is almost always the most expensive one. One over-powered choice looks harmless on its own, but multiplied across every prompt, agent, and workflow, and you get a big number on the P&L. All that, yet nobody chose which model on purpose. As we like to say, using a default is not a decision.

Migration projects fail in a predictable sequence. The technical work gets scoped. The timeline gets set. The engineering team starts moving workloads. Somewhere in the middle, dependencies surface that weren't in the original assessment, the double-run period extends beyond the budget allocated for it, and the project either stalls or completes at significantly higher cost than planned.

Cloud outage response got a brutal stress test in July 2026. In the span of nine days, three separate cloud infrastructure failures took large chunks of the internet offline: AWS CloudFront on July 16, Microsoft Azure West US on July 23, and AWS us-west-2 on July 24. None of them were caused by a dramatic data center fire or a nation state attack. They were routing faults, configuration translation bugs, and a piece of networking hardware on the path between a region and a metro area.

Cloud outage preparedness stopped being a nice-to-have this month. In a span of roughly 48 hours, Microsoft Azure lost a big chunk of its West US footprint and Amazon Web Services dropped connectivity between its us-west-2 region in Oregon and the Seattle metro. The AWS event alone rippled outward and knocked DoorDash, Reddit, Hulu, Apple Pay, Snapchat, Fortnite, and the PlayStation Network offline for millions of users, according to incident trackers. Neither outage was caused by anything exotic.