Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Question concerning zone redundancy and ACR traffic costs #788

Closed
MarkTopping opened this issue Nov 7, 2024 · 1 comment
Closed

Question concerning zone redundancy and ACR traffic costs #788

MarkTopping opened this issue Nov 7, 2024 · 1 comment
Assignees
Labels
feature-high-availability Issues related to geo-replication, zone redundancy and high availability question Qustions about features or clarifications needed triaged Use after the issue is triaged

Comments

@MarkTopping
Copy link

Hello,

I would like to understand and clarify whether or not customers incur network egress costs for ACR traffic when the ACR instance and the ACR consumer reside in different availability zones from one another.

There are a few common scenarios we could consider. In all cases though lets state that that both compute and the private endpoint for the ACR reside within the same VNet - thus no cross-vnet traffic.

  1. Suppose my compute is pinned to Availability Zone 1 and ACR is spread across AZ 1, 2, and 3 via the zone redundancy feature
  2. Suppose both my compute and ACR are spread across AZ 1, 2, and 3.
  3. Suppose my compute is pinned to AZ 1, and ACR has zone redundancy disabled but it got deployed into to AZ 2 (since we cannot chose the zone ourselves)

There is no mention of costs or implications to costs in the documentation concerned with ACR Zone redundancy found here: https://learn.microsoft.com/en-us/azure/container-registry/zone-redundancy. And there is no mention of cost impact when you use the Azure Portal to create your ACR instance.

However, there is a mention of costs in the ACR Best Practises page here: https://learn.microsoft.com/en-us/azure/container-registry/container-registry-best-practices

It states: "Pulling images from one datacenter to another adds network egress fees, in addition to the latency."

It seems to me that if a customer cannot choose a specific zone for their ACR to be provision in, and therefore cannot guarantee co-location next to their compute (if they are pinning their compute to a specific Zone); then it's simply down to luck whether or not their compute and ACR are co-located and therefore they would not be able to control their fees if that above statement holds true.

I'd also expect that if enabling zone-redundancy does amount to network egress costs being charges then this would be made clear in the page which documents the feature.

FWIW, and for interest, GitHub Co-Pilot is of the opinion that network egress charges are not applied for ACR traffic between datacentres - contradictory to the sentence in bold above.

Thanks.

@toddysm toddysm added question Qustions about features or clarifications needed triaged Use after the issue is triaged feature-high-availability Issues related to geo-replication, zone redundancy and high availability labels Jan 9, 2025
@johnsonshi
Copy link
Contributor

This is a very good question @MarkTopping. From the Azure Bandwidth pricing doc (https://azure.microsoft.com/en-us/pricing/details/bandwidth/), Azure does not charge for any fees within the same region even if they are from different Availability Zones within the same region. The Azure Bandwidth pricing page makes it clear that there are no additional egress charges in the FAQ section of the doc I linked.

Image

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
feature-high-availability Issues related to geo-replication, zone redundancy and high availability question Qustions about features or clarifications needed triaged Use after the issue is triaged
Projects
None yet
Development

No branches or pull requests

3 participants