Microsoft Azure Is Out Of Cores!

We’re not ones to kick a man when he’s down — but if you’re planning to drop your SOLIDWORKS PDM or DriveWorks estate onto a standard Azure VM right now, you should read the Reddit thread titled “Central_US out of V4 and V5 cores”. The upshot: Azure can — and does — run out of the exact VM/CPU core families teams count on. That matters a lot for engineering workloads.

What “out of cores” actually means

Cloud capacity isn't infinite. Every Azure region has a finite pool of physical server hosts and specific VM core families. When demand spikes or capacity is rebalanced, newer core families (the “v4/v5” style SKUs) may be temporarily unavailable. Practically that means you can’t provision the VM size you designed for, which bubbles up into missed deadlines, delayed migrations, or brittle workarounds.

Why this hits CAD/PDM teams harder

Engineering workloads aren’t commodity web stacks. Azure requires that:

• You need specific SKUs for predictable CPU, memory and networking behavior. Substituting a different series isn’t always acceptable.
• Storage/IO and network performance are critical for PDM archives and DriveWorks automation — you can’t just scale CPU and call it a day.
• Licensing and vendor support often expect validated architectures; if your VAR-approved SKU is unavailable, supportability can become murky.
• Geographic changes to solve capacity can introduce latency for engineers and complicate compliance.

Short-term fixes (what teams do when cores are unavailable)

• Choose an alternate SKU or region — but beware of latency and support tradeoffs.
• Reserve capacity with your Azure account team or create capacity reservations where possible.
• Use older, proven VM families if validated by your vendor, and size storage/IO accordingly.
• Engage Microsoft support/your account rep — they can sometimes approve a quota bump or mirror capacity.
• Consider Azure Dedicated Host or private cloud for guaranteed physical capacity.

A better approach for PDM / DriveWorks

Don’t treat Azure like a vending machine. For engineering workloads you need planning and guarantees, which EpiGrid's Private Cloud Host can provide.

• Pricing Consistency (dedicated hosts, or committed capacity through a partner like EpiGrid).
• Design for the workload — right storage tier, SQL architecture, application-consistent backup, and network topology.
• SOLIDWORKS Expects so vault administration can be handled by us.
• Use an engineering-first managed host if you can’t or won’t manage capacity guarantees yourself — EpiGrid holds reserved racks of validated architectures and handles monitoring and management of the system.

Bottom line

Azure running short of V4/V5 cores in a region is a reminder: cloud is powerful, but not all cloud-hosts are built the same. If you’re running PDM or DriveWorks, design for capacity or work with a partner who guarantees it. Otherwise you’ll trade instant provisioning for potential long outages and unhappy engineers — and nobody likes that.