Blog Post

Azure Data Factory Blog
1 MIN READ

Announcing public preview of Time-To-Live (TTL) in managed virtual network

lrtoyou1223's avatar
lrtoyou1223
Icon for Microsoft rankMicrosoft
Jun 28, 2022

Today we’re excited to share the public preview of Time-To-Live (TTL) in managed virtual network.

Background

Managed virtual network provides customers with a secure and manageable data integration solution. But due to the limitation of architecture, we need to provision computes in managed virtual network each time we execute an activity. This can lead to relatively long queue times. Especially when you have small jobs that are executed sequentially, it’s not very efficient. So we introduce a TTL feature that allows users to reserve computes and these computes won’t be released within TTL period after the last activity execution.

 

How does it work

You can configure TTL settings in the integration runtime creation or edit page if this integration runtime enables managed virtual network. Since we use different computes from interactive authoring, copy activity and pipeline/external activity, so you can set different TTL value respectively.

 

 

You can see more introductions and one demo in this video

 

 

Learn more about managed virtual network and TTL

Managed virtual network and managed private endpoints - Azure Data Factory | Microsoft Docs

Updated Jun 28, 2022
Version 1.0
  • SaiVijay's avatar
    SaiVijay
    Copper Contributor

    lrtoyou1223: If lot of copy activities are triggered in parallel from for-each loop, how does the queue time works for the parallel activities with this new TTL feature.

    I understand if copy activities are sequential, based on the TTL value set, the successive copy activity might use the reserved compute, but if activities are triggered in parallel, will it reserve multiple compute resources with the given TTL value? If yes, how does this impact the billing?