Major VM SKU Deprecations and New AKS Extension Rollout

· AI summary
919 new23 regressions20 parked
Daily executive summary

Compared with the 2026-07-08 snapshot, the 2026-07-09 scan recorded 919 new listings, 23 regressions, and 20 parked unknown transitions.

Representative changes: Standard Dc1s V2 VM size was no longer listed in 5 regions; Standard Dc2s V2 VM size was no longer listed in 5 regions; Standard Dc4s V2 VM size was no longer listed in 5 regions.

What changed: 15 first recorded disappearances, 8 recurring disappearances, 919 net-new regional listings.

What this means: Review the 23 regressions against existing regional targets and fallbacks first; treat the 919 new listings as options to validate. These are read-only catalog/list signals, not quota, capacity, or deployment results. Zero new delistings does not establish recovery of earlier missing listings.

7-day context (2026-07-03 to 2026-07-09): across 7 recorded scans, there were 2,639 new listings and 754 regressions. New listings led on 6 scans; regressions led on 1.

Archived commentary and representative evidence

Significant regional regressions were observed for Azure confidential compute VM SKUs (DCv2 series), with Australia East, UK South, UK West, West Europe, and West US all losing vmSkus.standard.dc1s.v2, dc2s.v2, dc4s.v2, and dc8.v2 from their catalogs. These SKUs had been continuously available for 58 days in most regions, with West US showing prior instability (1 disappearance). Coverage dropped from 9 to 4 regions, a 55.6% deprecated coverage rate. Still available fallback regions are australiasoutheast, eastus, northcentralus, and northeurope. For SREs, this means right-sizing, performance, and capacity fallback lists must be urgently rechecked. See VM size overview: https://learn.microsoft.com/azure/virtual-machines/sizes/overview

Azure AI model regressions also hit East Asia, which lost availability for aiModels.openai.gpt-5-2.2025-12-11, gpt-5-5.2026-04-24, and gpt-5.2025-08-07. These models are now only available in 33 of 64 monitored regions, with 7 remaining in Asia. For model deployment targets, latency, and data residency, review fallback model/version choices and placement options. Details: https://learn.microsoft.com/azure/ai-foundry/openai/concepts/models

In contrast, AKS extensionTypes.dcassociatesgroup.nats saw a major first-time rollout, now available in 58 regions including all monitored Oceania and Europe locations. This marks the first observed deployment for this extension, which enables managed cluster capabilities such as GitOps, policy, and observability via Azure Resource Manager-driven installation and lifecycle management. This broad expansion opens up new operational options for AKS clusters. More info: https://learn.microsoft.com/azure/aks/cluster-extensions

Belgium Central also gained Azure Functions Flex Consumption hosting plan and runtime support for .NET (6, 7, 8, 9, 10), Java (11, 17, 21), expanding serverless scale-out options with private networking and memory sizing. This is a regional expansion of existing signals, now covering 81.2% of monitored regions. SREs can now consider Belgium Central for burst scale and lower ops overhead. Details: https://learn.microsoft.com/azure/azure-functions/flex-consumption-plan

Overall, urgent action is required for VM SKU placement and capacity planning due to widespread deprecations, while new AKS and Functions capabilities offer expanded deployment choices.

Engineering context

  • australiaeast vmSkus.standard.dc1s.v2 available → unavailable deprecation candidate
    unavailable 0/58 prior days (0.0%)coverage 4/64 monitored regions (6.2%, -5 regions)coverage removed from 55.6% of prior regionsOceania: 1 current, 2 prior
    Still available in australiasoutheast, eastus, northcentralus, northeurope. VM size availability affects right-sizing, performance, cost, and capacity fallback choices. Azure VM sizes
  • australiaeast vmSkus.standard.dc2s.v2 available → unavailable deprecation candidate
    unavailable 0/58 prior days (0.0%)coverage 4/64 monitored regions (6.2%, -5 regions)coverage removed from 55.6% of prior regionsOceania: 1 current, 2 prior
    Still available in australiasoutheast, eastus, northcentralus, northeurope. VM size availability affects right-sizing, performance, cost, and capacity fallback choices. Azure VM sizes
  • australiaeast vmSkus.standard.dc4s.v2 available → unavailable deprecation candidate
    unavailable 0/58 prior days (0.0%)coverage 4/64 monitored regions (6.2%, -5 regions)coverage removed from 55.6% of prior regionsOceania: 1 current, 2 prior
    Still available in australiasoutheast, eastus, northcentralus, northeurope. VM size availability affects right-sizing, performance, cost, and capacity fallback choices. Azure VM sizes
  • australiaeast vmSkus.standard.dc8.v2 available → unavailable deprecation candidate
    unavailable 0/58 prior days (0.0%)coverage 4/64 monitored regions (6.2%, -5 regions)coverage removed from 55.6% of prior regionsOceania: 1 current, 2 prior
    Still available in australiasoutheast, eastus, northcentralus, northeurope. VM size availability affects right-sizing, performance, cost, and capacity fallback choices. Azure VM sizes
  • eastasia aiModels.openai.gpt-5-2.2025-12-11 available → unavailable deprecation candidate
    unavailable 0/58 prior days (0.0%)coverage 33/64 monitored regions (51.6%, -1 region)coverage removed from 2.9% of prior regionsAsia: 7 current, 8 prior
    Still available in australiaeast, brazilsouth, canadacentral, canadaeast, centralindia, centralus, and 27 more. Use this to evaluate model capabilities, regional deployment options, latency, and data residency. Azure OpenAI model availability
  • eastasia aiModels.openai.gpt-5-5.2026-04-24 available → unavailable deprecation candidate
    unavailable 0/58 prior days (0.0%)coverage 33/64 monitored regions (51.6%, -1 region)coverage removed from 2.9% of prior regionsAsia: 7 current, 8 prior
    Still available in australiaeast, brazilsouth, canadacentral, canadaeast, centralindia, centralus, and 27 more. Use this to evaluate model capabilities, regional deployment options, latency, and data residency. Azure OpenAI model availability
  • eastasia aiModels.openai.gpt-5.2025-08-07 available → unavailable deprecation candidate
    unavailable 0/58 prior days (0.0%)coverage 33/64 monitored regions (51.6%, -1 region)coverage removed from 2.9% of prior regionsAsia: 7 current, 8 prior
    Still available in australiaeast, brazilsouth, canadacentral, canadaeast, centralindia, centralus, and 27 more. Use this to evaluate model capabilities, regional deployment options, latency, and data residency. Azure OpenAI model availability
  • uksouth vmSkus.standard.dc1s.v2 available → unavailable deprecation candidate
    unavailable 0/58 prior days (0.0%)coverage 4/64 monitored regions (6.2%, -5 regions)coverage removed from 55.6% of prior regionsEurope: 1 current, 4 prior
    Still available in australiasoutheast, eastus, northcentralus, northeurope. VM size availability affects right-sizing, performance, cost, and capacity fallback choices. Azure VM sizes
  • uksouth vmSkus.standard.dc2s.v2 available → unavailable deprecation candidate
    unavailable 0/58 prior days (0.0%)coverage 4/64 monitored regions (6.2%, -5 regions)coverage removed from 55.6% of prior regionsEurope: 1 current, 4 prior
    Still available in australiasoutheast, eastus, northcentralus, northeurope. VM size availability affects right-sizing, performance, cost, and capacity fallback choices. Azure VM sizes
  • uksouth vmSkus.standard.dc4s.v2 available → unavailable deprecation candidate
    unavailable 0/58 prior days (0.0%)coverage 4/64 monitored regions (6.2%, -5 regions)coverage removed from 55.6% of prior regionsEurope: 1 current, 4 prior
    Still available in australiasoutheast, eastus, northcentralus, northeurope. VM size availability affects right-sizing, performance, cost, and capacity fallback choices. Azure VM sizes

Generated from the monitor's structured change evidence. See the status meanings for what each signal proves, or browse all posts.

Help make this page clearer

Two short answers; GitHub will open a draft for your review. No Forms, token, or automatic submission.

This repository is public. Review the text and image for private details. Opening the draft sends text/context to GitHub in its URL. Pasting or attaching an image uploads it to GitHub immediately.

Describe the desired outcome in plain text. This becomes the issue's ### Objective. After reviewing the issue, a maintainer can add azure-backlog to queue it for a scheduled agent; the website does not add that label or start a run.

Page context included in the draft

Nothing is posted by this website. Sign in to GitHub if needed, review, attach your screenshot, and choose Submit. Closing this panel clears the local draft and screenshot.