Ok .. so .. let me get one thing out of the way first: I’m not a Fabric expert. Far from it. I have a LOT to learn about Fabric myself 🙈. But this one is too important not to spread the word, so here’s my summary of what Kennie Pontoppidan showed in his Launch Edition video for 2026 release wave 2.
Mirroring from Business Central to Microsoft Fabric. Natively.
Coming in public preview in 29.x (Kennie says “likely 29.1”).
And yes .. I’ve been waiting for this for a long time. Believe me – I was one of the few that was haunting Microsoft every single occasion I got .. 😉 .. because in my view, this is the only sensible way to get BC data into Fabric: built into the product, supported by Microsoft, and not something you have to bolt on yourself and keep alive. Let’s be real .. copying data out of an ERP shouldn’t be every partner’s side project 😉.
As simple as 1 2 3
(Kennie’s words, not mine 😉)
- Create a mirrored database in Fabric (an “open mirroring” one).
- Tell Business Central where to put the data. In the new Fabric integration app, you plug in the workspace and the mirror database, hit “Connect to Fabric”, and BC sets up the plumbing. There’s a “Test connection” too.
- Pick your tables and your companies, click “Start synchronization“. That’s it.
You can also do step 3 with configuration packages. Microsoft has the list of tables its Power BI apps need, so that can simply go in a package, and ISVs can ship their own for whatever their solution needs. So no puzzling over which tables you need 🤷.
What do you get in Fabric?
Your tables show up in OneLake as Delta Parquet, the way AL sees them. Thanks to the SQL endpoint, you can query them with plain SQL, as if it were a normal database. Even I can do that 😅 (I think..).
How “live” is live?
I haven’t put this through its paces myself (yet 😉), so I’ll borrow from Stefano here: in his testing, the mirror runs a few minutes behind BC, depending on how volatile your data is. Deletes come across. You pick tables and companies, not fields .. everything in the table gets replicated. And so far nobody has run into table, row or volume limits. Oh, and the part I like most: none of this touches your API limits or your production database performance, because it’s not AL or APIs doing the copying.
Again .. that’s Stefano’s experience, not mine. Mine will follow once the sync app is on AppSource 🤷.
Monitoring and security
From what I saw in the video:
In BC, you get two new pages under a “Monitoring” menu: a synchronization overview (one line per sync run, basically) and synchronization details (what actually got pushed, table by table). Both come with ready-made views, so you can jump straight to the errors instead of scrolling through the happy runs. And since these are just pages, analysis mode works on them too .. instant mini-report.
Those same logs are also pushed to Fabric. So if your data team lives in Fabric and never opens BC, they can monitor the sync from their side, with the same SQL endpoint they use for the data itself.
What does it cost?
Let’s start with the easy part: on the BC side, nothing. No extra license, no consumption charge. Stefano Demiliani already wrote that down nicely in his post, so go read that one as well.
The Fabric side is where the money is. Disclaimer first: I’m reading the Fabric docs here, not the BC ones 😉. But since this all lands in an open mirroring database, this is how I understand the cost of mirroring in Fabric:
You need a running Fabric capacity. If you pause or delete it, nothing gets replicated.
The compute that replicates the data into OneLake is free. It doesn’t eat your capacity. (One footnote: Fabric’s “extended” mirroring capabilities, like Delta Change Data Feed and views, ARE billed. But that’s opt-in stuff on top, not the plain mirror)
Storage is free up to a limit: 1 TB of mirroring storage for every capacity unit you buy. Go beyond that, or pause your capacity, and you pay for OneLake storage.
Querying the data (SQL, Power BI, Spark) consumes your capacity at regular rates.
Fabric capacities come in “F” sizes (F2, F4, F8 … F64 and up), where the number is the amount of capacity units. So an F2, one of the smallest ones, already gives you 2 TB of free mirroring storage. Plenty for most BC databases, I’d think.
So in short: getting the data in is free, using it costs capacity. Check the Fabric pricing page for actual numbers, which depend on your region and SKU. And again .. if you’re a Fabric person and I got this wrong, ssshh 🤫 (kidding – comment below! ;-)).
And the Power BI apps?
All 9 Power BI apps get a Fabric-based version. Each one also comes as a multi-company version, with a company filter on top. They’ll be available “soon”, either as code samples or on AppSource.
So .. what now?
Watch the video. It’s only 10 minutes, and Kennie explains it way better than I can. Then keep an eye on aka.ms/bcanalytics. At the time of writing, the BC docs don’t cover the new mirroring yet, so more details are coming.
Kudos to Kennie and the team. And if you are a Fabric person: I’d love to learn from you. Drop a comment 😊.
Enjoy! I know quite a few of our customers will!
Loading ...