The alpha, twice a week
That was 3 min.
Get the best new moments in your inbox every Wednesday and Saturday. It's free.
Check your inbox
We sent a link to your address. Click it to finish signing up. Not there in a minute? Look in spam.
or
Latent Space · 3 min · Oct 28, 2026
Production database migrations risk downtime and split-brain data
The founders say the hard part of migrations is not rewriting code but avoiding downtime and split-brain data. They often favor landing with new workloads first rather than moving a production database outright.
Co-founder & CTO, Supabase
Shawn Wang (swyx), Host · Co-founder, AI Engineer
Wrote “The Rise of the AI Engineer” and founded Smol AI, maker of AI News.
About this segment
The founders say the hard part of migrations is not rewriting code but avoiding downtime and split-brain data. They often favor landing with new workloads first rather than moving a production database outright.
From “Why Agents Love Supabase — with Paul Copplestone and Ant Wilson”, Oct 28, 2026. Watch the full episode
Thanks. We'll look at and fix what's off.
“a minute of downtime means that you end up with basically dollars lost.”
Takeaways
- Even brief OLTP downtime can mean lost revenue.
- Running databases in parallel risks split-brain data.
- Teams often move new workloads before migrating production.
In this episode
Full episode- 41:29Supabase measures sales by incremental customer growth4:46
- 59:07Declarative schemas give agents a direct view of database state3:06
- 23:09A launch forced Supabase to migrate 800 databases in YCPro pick 3:01
- 9:02Multigres aims to prevent data loss during failovers3:44
- 12:45Turso helps Supabase scale down to billions of databases4:03
- 35:34Supabase set aside $1 million in stock for community members4:16
48:15Production database migrations risk downtime and split-brain data2:37
- 1:11:01Supabase spends about $300K a month on AI tools3:53Up next
Transcript
From the transcript Latent Space publishes with the episode.
- 48:28
I heard some crazy stat like that it might be different now 'cause they're trying to, Oracle are trying to move everything to the cloud. But I think it was like one year ago, someone said that only five percent of Oracle workloads by revenue ran in the cloud. And so, like almost everyone was saying, "Oh, let's rip out Oracle. Let's do migrations." People... That's what I mean, though. Yeah. Like, when I look through the history of database providers who have tried to run this strategy, it sounds amazing, and it's something that we'll of course try to do ourselves.
- 49:01
Why would we not try to do this? But it is incredibly difficult, and I think you would just wanna focus, first of all, on landing inside the enterprise. Uh, yeah, that makes sense. Yeah. It's, it's one of those things where, like, you know, when, when I s- see... When I talk to coding agent people all day long, the great promise is that you can finally migrate off of all this legacy software that you have. Yeah. Right? And, like, hopefully- That's true...
- 49:23
that, that means that you cut over everything. That's true, because one of the things that actually makes Oracle so sticky is, like, you end up writing these million-line stored procedures. Like, you stuff everything into your database. And so actually probably rewriting that is, these days, a lot easier. Right. Right. It's... So, like, hopefully coding agents are the boom. Uh, maybe they're not there yet.
- 49:43
Maybe you guys have to become more on-prem. Most of the problem around migrations is actually not, not rewriting it, and you can do migrations, for example, on the data analytics side quite easily from one, from one warehouse to another. The problem is that usually your OL- OLTP database is powering your most revenue-generating application. So a minute of downtime means that you end up with basically dollars lost.
- 50:08
And so then what you end up doing is having to run these two things in parallel, and then the risk of this is that you get this split brain where you've got some data running on one side, some on another, and if you try to reconcile s- split brain, anyone who's tried to do it, it's a very hard problem. So the risks are just so much higher when it comes to OLTP that a lot of the time people just start doing... splitting the workloads off, and they'll start new workloads.
- 50:33
They'll do a lot of operations here, and hopefully they'll migrate things over. Yeah. Usually then you've got two problems, though. [laughs] Well, speaking of OLTP, I mean, I, I...