Azure Virtual Machine HB-series are retiring on 31 August 2024

[azure-deprecation-automation]

Azure Virtual Machine HB-series are retiring on 31 August 2024

Deadline: Aug 31, 2024 Impacted Services:

• Azure Virtual Machine

More information:

• https://azure.microsoft.com/en-au/updates/hbseries-azure-virtual-machines-will-be-retired-by-31-august-2024/
• https://aka.ms/H-series-migration

Notice

Here's the official report from Microsoft:

As Azure HBv2 and HBv3 virtual machines have greater memory bandwidth, improved remote direct memory access (RDMA) networking capabilities, larger and faster local solid-state drives, and better cost/performance across a broad variety of HPC workloads, we are retiring our HB-series Azure Virtual Machine sizes on 31 August 2024.

After 31 August 2024, any remaining HB size virtual machines subscriptions will be set to a deallocated state, will stop working, and will no longer incur billing charges.

Note: This VM size retirement only impacts the VM sizes in the HB-series. This retirement announcement does not apply to the newer HBv2, HBv3 and HC series virtual machines.

Timeline

Phase	Date	Description
Announcement	Aug 18, 2021	Deprecation was announced
Deprecation	Aug 31, 2024	VMs will be set to a deallocated state, will stop working, and will no longer incur billing charge.

Impact

VMs will be set to a deallocated state, will stop working, and will no longer incur billing charge.

Required Action

Use the official migration guide.

Here's the official report from Microsoft:

We’re retiring HB VMs on 31 August 2024. All HB VMs will need to be de-allocated by that date. For recommendations on which VMs to migrate your HB workloads to or for the creation of net new VMs, please see this migration guide. Visit the Azure Virtual Machine pricing page for additional information.

Contact

You can get in touch through the following options:

• Contact the product group through email (email).
• Contact Azure support (link).

[azure-deprecation-automation]

This issue is automatically managed and suggest to use GitHub Discussions to discuss this deprecation.