FUAM : How can i have insight in sharing semantic models

[amien]

How can i see in FUAM lakehouse, who connects to a semantic model and how much CU that person is consuming.

I have checked FUAM_Lakehouse].[dbo].[activities] table .. do i need to look at the Activity or Operation column? ConnectFromApplications?

And how do i get to the CU? Because i don't see that in the activity table.

Thanks in advanced

[burakkaragoz · Accepted solution]

@amien90 , @amien ,

Thanks for clarifying — now I fully understand what you're trying to achieve.

You're essentially looking to track guest user activity (e.g. opening a shared semantic model via Power BI Desktop) and measure the CU consumption for that specific action.

Here’s the current state of things:

---

1️⃣ Tracking Guest vs Member Activity

• In the activities table (or similar logs), check for:
  • UserPrincipalName or UserId — guest users often show up with external domains or special suffixes (like #EXT#).
  • ConnectFromApplications — may show “PowerBI” or “PowerBI Desktop” if available.
  • Operation — look for values like Connect, Query, or ViewReports.

⚠️ Note: The logs don’t always explicitly say “guest”, so you may need to infer it based on domain or naming pattern.

---

2️⃣ CU Consumption for That Activity

• CU usage is not tracked per user or per operation — it’s aggregated at the workspace or item level.
• You can use capacity_metric_by_item_by_hour or similar tables to get CU usage for the semantic model.
• But there’s no direct join between user activity and CU usage.

---

🧪 Workaround (Approximation)

You could try this approach:

1. Use activities to identify when a guest user accessed the model.
2. Use capacity_metric_by_item_by_hour to get CU usage for that model during the same time window.
3. If only one user was active during that time, you can attribute the CU to them.
4. If multiple users were active, you can estimate CU per user based on activity frequency.

It’s not perfect, but it gives you directional insight.

---

Let me know if you want help writing a query to test this approach — happy to assist!

If my response resolved your query, kindly mark it as the Accepted Solution to assist others. Additionally, I would be grateful for a 'Kudos' if you found my response helpful.

*This response was supported by AI for translation and text editing purposes.

[Rufyda]

Hi
Glad that your query got resolved. Please continue using Fabric Community for any help regarding your queries.

[amien]

Thanks for the insight .. i understand what you mean .. unfortunately .. directional insight is not good enough when you want to backcharge certain operations

[burakkaragoz]

Hi @amien ,

Good question — tracking semantic model usage and CU consumption in FUAM can be a bit tricky since not all the info is in one place.

Here’s what you can check:

1. Who is connecting to the semantic model

You're on the right track with FUAM_Lakehouse.[dbo].[activities]. Look at:

• Operation column → should show things like Query, Connect, etc.
• ConnectFromApplications → helps identify if it's Power BI, Excel, etc.
• UserId or UserPrincipalName (if available) → shows who initiated the action

2. CU consumption

Unfortunately, CU usage is not directly logged in the activities table. To get CU-level insights:

• You need to look at the Capacity Metrics app (if you're using Power BI Premium or Fabric Capacity)
• Or query the Fabric Capacity Metrics APIs (if available in your tenant)

There’s also a resource_usage or capacity_utilization table in some setups — check if your Lakehouse has that. It might contain CU or memory/time-based metrics.

---

TL;DR:

• Use activities for user + operation tracking
• CU usage is tracked separately — not in activities
• You may need to combine logs from FUAM + Capacity Metrics for full picture

Let me know if you want help writing a query to extract specific usage patterns — happy to help.

If my response resolved your query, kindly mark it as the Accepted Solution to assist others. Additionally, I would be grateful for a 'Kudos' if you found my response helpful.

---

This response was supported by AI for translation and text editing.

[amien]

Thanks alot for your reply, i have a few follow-up questions:

• In Operation, i dont have Query, Connect .. i have ViewReports, ListDataAccessRoles, GetAppsAsAdmin, ConnectoGit .. that is the correct one right?
• Indeed i have ConnectFromApplications. But that could mean any application right? Excel, PowerBI Desktop. Nowhere is can see what application it is?
• I have several capacity_metric_by... tables : item_by_operation, item_kind_by_day etc. 
• It contains OperationName : Query, Render etc. But not UserID
• I have indeed access to Capacity Metrics
• But that ones doesnt contain UserID?
• I also don't see how i would join both sources. What is the key? 

I need to have the consumption of CU for each indivuel user and a specific operation (open the shared semantic model in PowerBI desktop)

[burakkaragoz]

@amien ,

Glad to hear the reply helped a bit!

I saw you mentioned you had a few follow-up questions — feel free to drop them here and I’ll do my best to help you out. If you’re looking into something specific like filtering by operation type or correlating activity with CU usage, just let me know what you’re trying to achieve and I can help you write a query or point you to the right resource.

Looking forward to your questions!

If my response resolved your query, kindly mark it as the Accepted Solution to assist others. Additionally, I would be grateful for a 'Kudos' if you found my response helpful.

*This response was supported by AI for translation and text editing purposes.

[amien90]

What i want to know: if guest users are connecting to a shared semantic model using powerbi desktop from our tentant and how much CU that costs.

for that i need to know 2 things:

1/ activity of the user; open shared semantic model using powerbi desktop (or other apps). I need the user to determine if its a member or a guest user

2/ how much CU was consumped for this specific activity

[burakkaragoz · Accepted solution]

@amien90 , @amien ,

Thanks for clarifying — now I fully understand what you're trying to achieve.

You're essentially looking to track guest user activity (e.g. opening a shared semantic model via Power BI Desktop) and measure the CU consumption for that specific action.

Here’s the current state of things:

---

1️⃣ Tracking Guest vs Member Activity

• In the activities table (or similar logs), check for:
  • UserPrincipalName or UserId — guest users often show up with external domains or special suffixes (like #EXT#).
  • ConnectFromApplications — may show “PowerBI” or “PowerBI Desktop” if available.
  • Operation — look for values like Connect, Query, or ViewReports.

⚠️ Note: The logs don’t always explicitly say “guest”, so you may need to infer it based on domain or naming pattern.

---

2️⃣ CU Consumption for That Activity

• CU usage is not tracked per user or per operation — it’s aggregated at the workspace or item level.
• You can use capacity_metric_by_item_by_hour or similar tables to get CU usage for the semantic model.
• But there’s no direct join between user activity and CU usage.

---

🧪 Workaround (Approximation)

You could try this approach:

1. Use activities to identify when a guest user accessed the model.
2. Use capacity_metric_by_item_by_hour to get CU usage for that model during the same time window.
3. If only one user was active during that time, you can attribute the CU to them.
4. If multiple users were active, you can estimate CU per user based on activity frequency.

It’s not perfect, but it gives you directional insight.

---

Let me know if you want help writing a query to test this approach — happy to assist!

If my response resolved your query, kindly mark it as the Accepted Solution to assist others. Additionally, I would be grateful for a 'Kudos' if you found my response helpful.

*This response was supported by AI for translation and text editing purposes.

[amien · Accepted solution]

BTW .. i'm going for this approach .. have you seen it?

https://pbi-guy.com/2024/04/24/how-to-extract-data-from-the-fabric-metrics-app-part-2/