Forum Discussion
Resolver IIISignificant capacity consumption increase since July 7th without major changes
Hi Fabric Community,
We are hoping to get some insights from the community regarding a significant and sustained capacity usage issue we've been experiencing since this past Monday. The Situation: Our Fabric capacity has been constantly hitting its limit. Upon investigating the Capacity Metrics app, we've identified a very specific cause: a sudden, fivefold (5x) increase in consumption from the operation category "OneLake other operations via redirect".
Our Architecture & Observations:
Our setup follows a Medallion architecture. The items primarily associated with this spike in usage are our "Platinum" Lakehouses. These Lakehouses serve our business layer and consist exclusively of shortcuts that point to tables in our "Gold" Lakehouse. We are wondering about two recent changes and if they might be related:
- Shortcuts: Given that the affected items are almost entirely composed of shortcuts, we are wondering if there has been a recent change in how usage for shortcuts is calculated or billed.
- OneLake Security: The source "Gold" Lakehouse was recently rebuilt to enable OneLake security. The "Platinum" Lakehouses that consume from it were then updated to point to this new, secured source.
Our Questions to the Community:
- Has anyone else observed a similar, sudden increase in CU consumption under the "OneLake other operations via redirect" category, particularly when using shortcuts?
- Is there any official documentation or known change regarding the cost model for accessing data through shortcuts that could explain this?
- Could the introduction of OneLake security on the source data be a contributing factor to how consumption is billed for these "redirect" operations?
We would be very grateful for any insights, similar experiences, or guidance you might have. Thank you in advance!
To provide more context, here are two graphs from our Capacity Metrics data showing the change.
Graph 1: Capacity Usage per Item As you can see, our primary Platinum Lakehouse shows a clear fivefold increase in CU consumption over the observed timeframe.
Graph 2: Capacity Usage per Category This graph shows that "OneLake other operations via redirect" has become our top capacity consumer by a significant margin, which was not the case before this week.
Hi ObungiNiels ,
Thank you for reaching out to the Microsoft Fabric Community Forum.
To address your questions, we can confirm that while there is no available documentation indicating a change in the cost model for shortcut-based access, we have observed some community feedback suggesting that using shortcuts and secured sources may lead to higher CU consumption. In your situation, enabling OneLake security on the Gold Lakehouse could be contributing to the increased usage, especially if the shortcut access now involves extra security enforcement layers or redirect logic. These actions may require additional backend compute or more thorough access checks, which might appear under the “redirect” category in the Capacity Metrics app.
Regarding your question about whether others have noticed similar behavior, while we can't speak for every tenant, there have been cases where customers using cross-workspace shortcuts or adding security features like ACLs or row-level access have reported differences in CU attribution.
To proceed efficiently, we recommend reviewing whether there have been any recent changes to your workloads, such as newly scheduled pipelines or queries, which might have increased access frequency to the Platinum Lakehouses.
Please also confirm if shortcuts are being utilized across workspaces and whether row-level security or OneLake access controls are implemented, as this information could assist in further isolating the root cause. Enabling detailed Capacity Metrics logging, if not already active, is also advised to help identify specific users or operations affecting consumption patterns.
For further clarity, Microsoft’s official documentation on OneLake consumption explains how CU usage is attributed, especially when using shortcuts and security features. It confirms that:
- CU consumption is attributed to the workspace accessing the shortcut, not the source.
- Security features like RLS can increase CU usage depending on the number of rows secured.
If the issue persists without any workload changes, we recommend submitting a formal support ticket with Microsoft to initiate a backend review by the engineering team.
Please keep us informed of any further insights or updates on your end. I hope my suggestions give you good ideas, if you need any further assistance, feel free to reach out.
Thank you.
4 Replies
Hi ObungiNiels ,
Thank you for reaching out to the Microsoft Fabric Community Forum.
To address your questions, we can confirm that while there is no available documentation indicating a change in the cost model for shortcut-based access, we have observed some community feedback suggesting that using shortcuts and secured sources may lead to higher CU consumption. In your situation, enabling OneLake security on the Gold Lakehouse could be contributing to the increased usage, especially if the shortcut access now involves extra security enforcement layers or redirect logic. These actions may require additional backend compute or more thorough access checks, which might appear under the “redirect” category in the Capacity Metrics app.
Regarding your question about whether others have noticed similar behavior, while we can't speak for every tenant, there have been cases where customers using cross-workspace shortcuts or adding security features like ACLs or row-level access have reported differences in CU attribution.
To proceed efficiently, we recommend reviewing whether there have been any recent changes to your workloads, such as newly scheduled pipelines or queries, which might have increased access frequency to the Platinum Lakehouses.
Please also confirm if shortcuts are being utilized across workspaces and whether row-level security or OneLake access controls are implemented, as this information could assist in further isolating the root cause. Enabling detailed Capacity Metrics logging, if not already active, is also advised to help identify specific users or operations affecting consumption patterns.
For further clarity, Microsoft’s official documentation on OneLake consumption explains how CU usage is attributed, especially when using shortcuts and security features. It confirms that:
- CU consumption is attributed to the workspace accessing the shortcut, not the source.
- Security features like RLS can increase CU usage depending on the number of rows secured.
If the issue persists without any workload changes, we recommend submitting a formal support ticket with Microsoft to initiate a backend review by the engineering team.
Please keep us informed of any further insights or updates on your end. I hope my suggestions give you good ideas, if you need any further assistance, feel free to reach out.
Thank you.
Hi ObungiNiels ,
I wanted to follow up on our previous suggestions. We would like to hear back from you to ensure we can assist you further.
Thank you.
Hi ObungiNiels ,
I wanted to check if you had the opportunity to review the information provided. Please feel free to contact us if you have any further questions.
Thank you.
- ObungiNiels
Hi there,
we have seen the capacity gradually descrease in the days following the spike. We have not heard any other occurrences of such spikes and are still happy if users can share insights on capacity consumption caused by shortcut tables.
For now, I think we can close this topic.
Kind regards,
Niels
