Xenonnskie Found a closely-matching thread from a month earlier that may be the same underlying issue: Fresh Hobby deployments return NOT_FOUND after "Increased invocation failures" incident (July 17, 2026). The symptoms line up closely with what we're seeing: Deployment metadata shows readyState: READY, readySubstate: PROMOTED — everything looks healthy But requests never reach the function/container — zero runtime logs, in our case confirmed even via a 5-minute runtime-logs API query Older deployments continue working normally while fresh builds of the identical commit fail That thread doesn't show a c

Amy Egan Thanks for the feedback! I shared this with the AI team so we can make sure the agent gives correct info

Amy Egan Thanks for making suggestions. I've shared this with the team 😃

Amy Egan Thanks for reporting. I shared with the team to get clarification on these numbers

Zautomation Real That is a difficult trade-off. $350 per month is significant, but a migration can also become a substantial hit once you account for the engineering time, operational changes and risk involved. I dealt with a similar situation before, and one thing that helped was establishing exactly what would need to move before estimating the migration. With PHI, that scope can be wider than it first appears. It is not only about where the database or backend runs. PHI can also surface through requests, logs, analytics, caches, previews and error reporting. Mapping those paths gave us a much clearer pictur

Session Sos I'm not sure what point you're trying to make; I have a clear path for migrating elsewhere. The issue is that Vercel would rather loose customers than make available a certification that is generic and free on most platforms. An elevated recurring fee for a one time cert is beyond ridiculous.

Ryu Hi Mohsin, Good catch. That definitely looks like a diagnostics/UI bug rather than an actionable domain error. If the real issue was DNS misconfiguration, then showing a value like: is pretty misleading, because it sends people toward their app/runtime stack instead of the actual domain/certificate checks. For anyone else landing here, I would not treat that kind of string as the cause of the cert failure. I’d still debug the domain normally: live A/CNAME, no conflicting AAAA, CAA records, and whether the HTTP validation path can reach Vercel cleanly. The useful bug report details here are pro

Mhslr This just happened to me. I removed and added back the domain to fix it.

Anshuman Bhardwaj Hi @amitpatel, I've shared this feedback with our team. I'll keep you posted here.

nerdynikhil If the Vercel team is open to it, I'd be happy to contribute the implementation to the repo.

galb An open-source CLI does most of this today. I made it: vercel-analytics-api (MIT, unofficial). It uses your vercel login, so it needs no new token. My tests on a real team: npx vercel-analytics-api visitors --since 30d - visitors and page views per project npx vercel-analytics-api top referrers -p my-app - top referrers as a table npx vercel-analytics-api series -p my-app --json - visitors per day as JSON UTM breakdowns need Web Analytics Plus.