Fabric June 25 release - GA features

 As we head into July, we get what is likely to be our last major update until Fabcon Vienna. With holiday season and Microsoft's end of financial year, we don't typically see July and August releases, and I wouldn't expect any different this year.

For those new to the blog, I post three separate blogs covering Fabric GA features, Power BI GA features, and preview features both combined. This one looks at the release notes that aren't clearly preview (though don't expect massive changes this month as we head into what are typically quieter times for Microsoft).

First up is data engineering.

Data engineering

Notebook version history

On the surface this one seems a minor change, but actually it's a massively handy feature. At one time or another, pretty much everyone makes a change that they need to undo but they haven't committed changes to version control as it's code they're still working on and isn't ready for a commit.

Until now that meant, in Fabric, that if you made a change you wanted to undo then you'd be out of luck if undo had lost the history. With this feature, this becomes less of an issue. Checkpoints are automatically created every 5 minutes or you can manually create checkpoints when you need.

All checkpoints are kept for up to a month.

Monitor improvement for T-SQL notebooks

This one is hidden in an area full of preview features and I actually missed it first time around.

For those using notebooks against a Warehouse or SQL endpoint, you are now able to access a run history for running, succeeded, cancelled and failed queries over the last 30 days.

New notebook creation dialog

Previously when you created a new notebook a default name was given and you had to go and rename the notebook. Now we get a pop up to create the notebook with the desired name and, for those using task flows, you can assign it to a stage at the point of creation.

Not a major update, but one that will improve the UX for certain.

RTI (Real time intelligence)

Azure data monitor in KQL Queryset

Previously you had to go through a specific steps to be able to add activity monitor data into the results set. 

Whilst this isn't new functionality, it certainly reduces the effort and makes it easier for people to use this data within KQL.

Sharing KQL queries

Again another UI redesign, nothing major in terms of new functionality.

Copilot KQL support

Now you can have Copilot write KQL queries within real-time dashboards.

Again, not a massive upgrade, but useful for those using Copilot. As with all Copilot features, if you are in the UK and decide to enable it please be aware that your data could go to the US in some situations (and it isn't clear in the Microsoft documentation).

SQL transform operator

Up until now, we'd been limited to the operators that could be used within an eventflow to transform the data. Now we are able to write custom SQL to manipulate event data in flight.

As you'd expect this comes with IntelliSense, auto complete, and preview functionality. Whilst the event stream supported a lot of the typical manipulations you might want to do, having this level of flexibility is definitely going to be helpful in the future.

Multiple schema support

This is one that really should have been a basic functionality when schema support was added, but at least it's sorted now.

Confluent schema registry support

For those using Confluent schema registry, we now get support for that in Fabric. If this applies to you, go and check out the blog.

Private end point support

Another one that ideally should have come as part of the MVP, but those using managed private endpoints on event hubs and IoT hubs can now ingest data into a Fabric Eventstream.

Data Factory

Azure data Factory item in Fabric

The ability to mount existing ADF pipelines into Fabric is now GA. For those looking to migrate existing Synapse setups into Fabric, this has just made the job a lot easier.

As expected, all pretty light this month. Till next time.