Backup Appliance
You can choose the size of the Azure VM running Veeam Backup for Microsoft Azure during the deployment, or change it later as the environment grows.
|
Note |
|
In Veeam Backup for Microsoft Azure version 8.1, you can only choose the B2ms, F4s_v2 or F8s_v2 VM size. |
General Recommendations
The following recommendations and examples apply to the latest Veeam Backup for Microsoft Azure builds.
|
Azure VM Size |
Recommended Maximum Number of Protected Workloads |
Recommended Maximum Number of Launched Worker Instances |
|---|---|---|
|
B2ms (2 vCPU, 8 GB RAM) Note: The minimum amount of system RAM recommended to perform Virtual Network configuration backup is 8 GB. If you plan to protect Azure Virtual Network configuration components, it is not recommended that you choose the B2s Azure VM size — consider choosing larger sizes. |
|
20 |
|
F4s_v2 (4 vCPU, 8 GB RAM) |
|
150 |
|
F8s_v2 (8 vCPU, 16 GB RAM) |
|
300 |
|
Note |
|
For product deployments running on Azure VMs whose size is larger than D4s_v3, Veeam Backup for Microsoft Azure may simultaneously launch 100 worker instances per region — however, this can trigger throttling issues in Microsoft Azure. If you are facing these issues, it is recommended that you use a maximum of 70 worker instances per region at a time. Alternatively, consider reducing the number of worker instances launched simultaneously by configuring different schedules for your backup policies or by specifying different target regions. For more information, see Performing Backup. |
Veeam Backup & Replication Integration
When you connect a backup appliance to the backup infrastructure, its backup policies, cloud-native snapshots, image-level backups, backup repositories and sessions are imported into the Veeam Backup & Replication database.
You can connect multiple backup appliances to a single Veeam Backup & Replication server. However, when working in an Azure subscription with cross-region data transfer, it is recommended that you use one Veeam Backup & Replication server per region, to help you avoid latency issues and meet potential data residency regulations.