There's no metric available for Azure Analysis Services to show the percent utilisation of QPU - only absolute QPU usage.

This makes it hard to alert on because a single threshold value won't scale with the size of the instance.

Azure SQL offers a DTU percent consumption metric which is a function of usage/DTU limit of the instance.

Is there a good reason the equivalent can't be added for AAS?

I'd like to see XMLA support for Azure Analysis Services the same way there is for Power BI.

If I have a Azure Analysis Server/Datbase I should be able to query it via XMLA

Right now you need to have an account in the same AD as the azure AS subscription tenant in order to be able to use your gateway for AS, which isn't convenient as you would need a separate account for every customer, which defeats the whole purpose of having 1 singe guest account to use.

Store Tabular Model "Views" that Users can see similar to tables. I have developers who write custom dax queries. These queries rely on the security model to the right information to each end user. This is like an advanced calculated table that does not execute until the query is sent. Currently, calculated tables calculate during model processing meaning user specific scope is not available. This would be similar to views in traditional SQL databases but just be predefined dax saved to the model definition.

To provide the information regarding how to customise the upload time in Azure Analysis Services. It seems the default upload time is set to around 4 min. What if we need to upload bigger files?

The AAS web designer preview is out for almost 2 years, yet it still supports only 3 data sources when creating a new data model. This is unusable in most scenarios. This is especially ridiculous when there is Power Query built-in already in AAS. As a starter, a simple data connectors like "Common Data Service", "Power BI Dataflows" or "M Script" would be sufficient for getting past that weird limitation.

Spark Connector for Azure Analysis Services doesn't seem to be working. I attempted different connstr including the one used successfully in Power BI. Nothing seems to work. No documentation either.

we have setup aas with a link protocol and all users have dual factor authentication as per https://community.powerbi.com/t5/Issues/Azure-Analysis-Service-live-connection-issue-with-a-pbit/idi-p/550868 .

In powerBI if you use the link:// protocol to connect you have to go through the dual factor twice on a pbix and on a pbit (template) when you reopen it it throws up 12 dual factor auth boxes and will return an error. Not sure if the problem is power bi or whether it is the client libs but raising here. Using Oct 2018 PBID and 15.0.2.39 client library.

currently if you write a query in DAX, and you run it, you can see the result
BUT if you switch to the MODEL view, and then go back to the DAX query tab, everything shown previously is reset (it would be great to keep it). it's like if when you go back to that interface you are actually refreshing the page

possibility of querying tables of the model in t-sql besides dax

Once Ia DAX query has been written (or generated) it would be nice to have a SAVE QUERY, so that for testing data within the model it's easier to rerun

Perhaps I am doing something wrong but in Log Analytics there is a Duration_s column that provides the length of time of a query_end event but as it is a string (_s) it is not available to be added to the x axis.

This seems like a pretty basic requirement of Log Analytics to chart duration. Why was duration created as a string when surely it is really numeric?

https://docs.microsoft.com/en-us/azure/analysis-services/analysis-services-server-alias indicates that the link:// protocol can only be used with Azure Analysis Services but this (combined with the fact that we can't automate the changing of connection strings in PowerBI.com means that a migration from onprem to PAS is a one-way operation which makes the change boards very uncomfortable.

Please allow this to point to both onprem and PAS SSAS instances

Access to Azure SSAS servers can be aliased using link:// in Power BI Desktop, thus allowing admins to load balance with Traffic manager and move models without redeployment. However this approach does not work with with on-premise SSAS servers since they do not present an https endpoint by default and is accessed by the cloud service via the Azure Gateway. Could you please provide an option for load bursting or even just failover from on-prem to Azure?

We have found that when viewing a cube pivot based on AAS that all data values are imported as string, regardless of the type set in the model.
This is in spite of the fact that when viewing standard Tabular Analysis Services models in Excel data types are respected.
Recognising data types allows for the creation of date hierarchies and mitigates sorting issues that occur when sorting by text values.
Please import data into the correct types when an AAS model is being viewed in Excel.

There is no copy and past option in Diagram /Grid view to copy and paste the table. Sometimes reporting team requires same copy of existing table for achieving the functionality of Chart As. In this case copy and past is very useful

It would be useful if it were possible to change the server name in Analysis Services by changing the connection for the cube.

It's inconvenient that we have to recreate the server to change the name.

Right now when you view the model as a script if you copy and paste this to use it will be decorated with "" so it's not valid JSON.

Would be better to have an export button so we can easily use the script.

Fix or at least provide an option to suppress showing of blanks where there are no blanks in the relevant table column in client tools (Pivot Table, Power BI)