Fabric Is Just Plain Unreliable, and Microsoft's Hiding It • Hacker News Board

Re: Fabric Is Just Plain Unreliable, and Microsoft's Hiding It

#11

Post by rcarmo » Tue, May 20, 2025, 7:09 AM UTC

flexagoon wrote: post #10

Just in case someone else also had no idea what Microsoft Fabric is: it seems to be some sort of integrated platform for data processing and analysis with AI features (because of course there's AI)

It’s way more than that. It’s an end to end data integration, data lake and pipeline management platform.

Re: Fabric Is Just Plain Unreliable, and Microsoft's Hiding It

#12

Post by flexagoon » Tue, May 20, 2025, 9:50 AM UTC

rcarmo wrote: post #11

flexagoon wrote: post #10

Just in case someone else also had no idea what Microsoft Fabric is: it seems to be some sort of integrated platform for data processing and analysis with AI features (because of course there's AI)

It’s way more than that. It’s an end to end data integration, data lake and pipeline management platform.

Sure. I have no idea what any of those words mean though because I've never worked with a platform like that. The way I described it is how I understood it as a person who has no idea what it is.

Re: Fabric Is Just Plain Unreliable, and Microsoft's Hiding It

#13

Post by datadrivenangel » Wed, May 21, 2025, 1:15 PM UTC

dexwiz wrote: post #8

I wonder if this is a difference in mindset of OLAP vs OLTP. I have worked with OLAP teams where they were dropping millions of records, and the solution was usually just restream the data and shrug. But if any service backed by OLTP was dropping writes, its all-hands-on-deck incident with detailed postmortems and exec meetings.

If you don't own the source system, then it's much harder to be incentivized to care.

Also the better data teams do postmortems and exec meetings.