What happened
An SEO reported that Search Console's 'Request a recrawl' function for robots.txt was returning an Unknown error — and across several properties, not just one. John Mueller then confirmed on Bluesky that he hits the same failure on some of his own sites. That's about as clean a your-side/their-side verdict as these things get: when Google's own John Mueller can reproduce the bug, the bug is Google's.
It hasn't been confirmed as a broad, platform-wide outage. The pattern so far is intermittent and partial — some properties fail, others may not — with no timeline for a fix and no acknowledgment beyond Mueller's confirmation that the error is real.
Why this matters
The damage from a broken button isn't the button — it's what people do next. The realistic sequence: you edit robots.txt during a redesign or to unblock a section, hit the recrawl request to hurry Google along, get an error, and conclude your edit broke something. So you edit again. Robots.txt is one of the very few files on a store where a nervous second edit can knock out crawl access to the entire catalog.
The button was always a convenience, never the mechanism. Google refetches robots.txt on its own cadence regardless — typically within a day — so a failed request delays nothing that wasn't already scheduled. Your file gets picked up either way. The stakes are real for a store, though: robots.txt gates whether Googlebot can reach your categories, products, and faceted URLs at all, which is exactly why the file deserves calm handling and the recrawl button doesn't deserve your trust as a deadline tool.
Worth clocking the pattern, too: this lands in the same month the page-indexing report kept dropping from daily to weekly data. Two Search Console defects running at once is a reminder that the console reports on your site — it isn't your site. When it misbehaves, verify against primary sources: your live files, your server, your logs.
What to do about it
Verify the live file yourself
Load yourdomain.com/robots.txt in a browser and confirm the edit is live. Then open Search Console's robots.txt report to see the exact version Google last fetched and when. Those two checks answer everything the broken button would have.
Log the error and stop
Note the property, the timestamp, and the error, then leave the file alone. Re-editing a correct file in response to a Google-side bug is how correct files stop being correct.
Plan robots edits around the natural refresh
Unblocking a section for a launch? Make the edit and expect pickup on Google's normal cadence — usually within a day — rather than counting on the button to force it. Build that day into the launch timeline instead of treating the recrawl request as a guarantee.
Put robots.txt under version control
Keep the file in git, or at minimum a dated changelog. When something goes sideways weeks later, proving exactly what was live on which date turns a guessing game into a lookup — and makes a panic edit trivially reversible.