Post

Apparently someone from Microsoft told my client that Azure integration services is the best practice way of ingesting data into Fabric. My client is now confused why we're using things like mirroring and notebooks instead. ...I'm looking forward to the meeting with that Microsoft employee 🙃

22/9/26, 4:41 pm1 repost20 likes

Justin Bird 🏴󠁧󠁢󠁷󠁬󠁳󠁿

@justinjbird.com

Ok, but what is Azure Integration Ser…OH FK NO!!!

Steph Locke

@stephlocke.bsky.social

That's someone sticking with the product "families" from the marketing side. logic apps, functions, data factory etc sit in that bucket

Justin Bird 🏴󠁧󠁢󠁷󠁬󠁳󠁿

@justinjbird.com

That is fair, my skim read didn’t get past “Unify systems with AI-powered integration”

Steph Locke

@stephlocke.bsky.social

I've taken it all with a pinch of salt since 10 years ago when a Microsoft page referred to etl as "a solved problem" 😂

RyanB

@ryanbseattle.bsky.social

Give 'em hell C.

Dr Greg Low

@greglow.bsky.social

Mirroring isn't a great solution, at least until it provides an easy way to determine which data has changed. It could be a good solution, with some enhancements. Until then, it's suitable for people building DirectLake solutions directly on transactional data structures (bad) or toy sized systems.

Dr Greg Low

@greglow.bsky.social

Might also work somewhat if the source was already an analytic system but they rarely are. Adding a few system generated columns would make it much better for incremental loading, which you need for any sizable system where the data needs transformation.