Databases should contain their own Metadata – Use SQL Everywhere

A startup called Floe says your database should tell you about itself — who’s using it, what’s slow, what’s huge — all answerable with plain old SQL. Their pitch: put all the “data about your data” (metadata) inside the database as queryable system tables, so you don’t need extra dashboards or logs. They even crack a pun about “table stakes,” and tease a beta, GitHub in tow.

The crowd? Immediately split. One camp rolled its eyes: veterans asked if this is just the same “information schema” every major database already has. Another camp pushed back, saying Floe isn’t just listing tables — it’s making everything queryable: stats, lineage (where data came from), freshness, access patterns, even cost and latency. Translation: not just “what tables exist,” but “who touched what, when, and how much it cost” — all with SQL. Meanwhile, a drive‑by heckler yelled “AI slop,” because it’s the internet.

The drama hits that familiar nerve: innovation or marketing? Fans want fewer dashboards, more answers in one place. Skeptics smell a rebrand of features databases have had for years. Bonus meme: the post’s quip about “suffering Grafana together” turned into a running joke, with commenters picturing a world where your database admits, “It’s me. Hi. I’m the problem, it’s me.”