We don’t usually cover headlines here, but two things happened this month that are worth a straight read rather than a press-release summary: Azure’s growth numbers, and a real outage. One is good news dressed up as good news. The other is bad news that’s actually useful if you take the right lesson from it.
Azure’s growth keeps compounding
Microsoft’s fiscal Q4 results put Azure revenue growth at 43% year-on-year, the fastest quarterly growth since early 2022, and Azure has now passed $100 billion in annual revenue for the first time. The contracted backlog sits at $678 billion, with a further $329.1 billion in signed-but-not-yet-started data centre capacity (Bloomberg, Axios).
What that means practically: Microsoft isn’t slowing down on capacity or feature investment, so the platform you’re building on today keeps getting more capable underneath you without you doing anything. It’s also a reminder that demand (and by extension, contention for capacity in popular regions) is only going up. If your business is planning a significant Azure expansion, region and SKU availability is worth checking early rather than assuming capacity is unlimited.
The July 23 outage was a process failure, not a platform failure
On July 23, a maintenance automation process in Microsoft’s West US region removed critical network routes, causing a multi-hour disruption to Azure and Microsoft 365 services in that region (w.media).
We’re not going to pretend this doesn’t happen, and we’re not going to pretend it’s uniquely an Azure problem either, every major cloud has had a version of this. What’s actually worth taking from it: this was caused by automation making a change it shouldn’t have, in a single region, which is exactly the failure mode that region redundancy and properly scoped change management exist to protect against. If a single region going down for a few hours would have been a serious problem for your business, that’s not an argument against Azure, it’s an argument for reviewing your resilience architecture. That’s precisely the kind of gap a Support Pack engagement is built to catch before it becomes an incident report.
What we’re actually watching
Growth numbers are good context, but they don’t change what we tell clients week to week. The outage is more useful: it’s a concrete, recent example to test your own setup against. If you’re not sure whether a regional issue like this would have taken you down too, that’s a five-minute conversation worth having before it’s a five-hour one.
If you want a second opinion on how your Azure environment would hold up, get in touch.