On July 17, 2026, Supabase reported a bigger problem starting at 18:17 UTC. It lasted about two hours and was resolved at 20:22 UTC. Supabase has since confirmed the disruption is over.
Who this kind of outage hits
If your app or workflows run on Supabase, a two-hour bigger problem means your database calls, auth, or storage may have stopped working, quietly, with no alarm going off on your side. The people who tend to find out first are not the builders. They are the customers who tried to log in, save something, or load a page and got nothing. By the time a customer writes in to say “your app is broken,” the outage may already be over, but the damage to trust is done. That gap, between when the tool stopped and when you heard about it, is the whole problem.
Why this is especially rough without an engineering team
A solo operator or a small team has no logs to check, no error screen to read, no alert firing in a Slack channel. The work just stops moving. A workflow that should have run at 18:30 UTC simply did not run. An order did not get processed. A user could not sign up. Nothing told you. The first signal was a confused or frustrated customer, and by then you are already behind, apologizing instead of explaining.
That is not a failure of preparation. It is just how these tools work when something goes wrong on their side. The status page updated, but you were not watching it.
Timeline
- 18:17 UTC, July 17, 2026: Supabase reports a bigger problem begins.
- About 2 hours of disruption.
- 20:22 UTC, July 17, 2026: Supabase reports the issue is resolved.
How a watcher catches this before your users do
NoCrash reads Supabase’s public status page every minute. The moment that page flips from working to having trouble, NoCrash sends you a plain-language heads-up, in words you can act on, without you having to go check anything. That means you find out within a minute of Supabase’s own report, not an hour later when a customer emails you.
It also watches the things you ship. If you have n8n workflows, NoCrash watches those too, so a quiet stall on your own side surfaces the same way. If you give it a URL or add a small JS snippet to your app, it watches that as well. The goal is one place where both the tools you build on and the things you ship are visible together, in plain English, before your users notice something is wrong.
NoCrash does not find the outage before Supabase’s own status page does. It reads that page and tells you fast. That is the honest version of what it does, and for most operators it is exactly what was missing here.
The authoritative account
For the full official record of this outage, go to Supabase’s status page: https://stspg.io/xf7wf1421m7x