atahf Hi Vercel Team! I'm having the same problem, can someone look into this? Team ID: team_H8Rsw11fVMF2qEUaPptIUW2q ps. I had a quick look in community posts and this issue seem to be happening quite often for most of Hobby plans

spanishkangaroo Hi! Same from my side. I just added the credit card today. Everything seems ok and I even got the 5$

Pauline P. Narvas Where is it in the docs? Can you share a link?

Ward Van Laer “In push mode, you can set a maximum concurrency per consumer group. This limits how many messages can be in-flight simultaneously for that group. If the limit is reached, Vercel holds back delivery until an in-flight message is acknowledged or its lease expires.“ https://vercel.com/docs/queues/concepts

Vicent Galiana We are encountering this limitation in a production-like workload using `@vercel/queue` 0.5.0 and a JavaScript `queue/v2beta` push consumer. One publish operation created 110 messages. Vercel started all 110 first-delivery callback invocations within 18.354 seconds. The consumer calls a rate-limited downstream GraphQL service, and 49 deliveries failed with HTTP 429 and required retries. We need to configure the push consumer group with a maximum concurrency of `1`. The documentation states: > “In push mode, you can set a maximum concurrency per consumer group.” However, we cannot find a supported configuration surface: * The consumer group under **Project → Observability → Queues** is read-only. * The documented `queue/v2beta` trigger schema exposes `topic`, `retryAfterSeconds`, and `initialDelaySeconds`, but no `maxConcurrency`. * `@vercel/queue` does not expose a push consumer-group update method. * `Vqs-Max-Concurrency` appears to apply only to poll-mode receive requests. Could you please confirm: 1. Is push-mode maximum concurrency currently available? 2. If it is internally configurable, can Vercel set our consumer group limit to `1`? 3. Is there a planned supported configuration through `vercel.json`, the dashboard, or an API?

Murderteeth +1. We’re on the Pro plan and need API tokens for monitoring/automation that only requires read access. The lack of read-only permissions prevents us from using API tokens for these use cases. Would love to see read-only or permission-scoped API tokens made available on Pro, ASAP.

Belty6979 5979 This is a really interesting use case. A native iOS client for managing deployments and checking project status would be genuinely useful, especially when you’re away from a laptop. The OAuth + PKCE approach also sounds like a much better experience than asking users to manage API tokens manually. Hope Vercel opens up the permissions beta soon.

Amy Egan Great blog post, as always! Looking forward to learning how it goes with your internal apps

Jonoroboto Don't worry Amy, it's guaranteed we'll break at least a core part of it, trying to fix security exploits 😂

Belty6979 5979 That’s a really interesting use case for Deepsec. Finding issues on a few hundred pages is already impressive, so I’d be curious to see what it uncovers when you run it against your internal apps.

Cris Precisely what's pushing away from adopting a gateway towards using bedrock directly. There are reasons why I would prefer AI Gateway, but data residency is a strong requirement and a big marketing point for many companies.

Belty6979 5979 I can definitely see why this would be important for teams with strict compliance requirements. Having a clear region setting would make the AI Gateway much easier to consider for those use cases.

Belty6979 5979 Yeah, I can see why this would be a concern for companies handling sensitive data. Having a simple way to choose the region would definitely make the gateway more useful.

ADALIGO Thanks for sharing this **Amy Egan** 🙌 Super helpful guide. I’ve only been here a few days but already learned a lot — the community is really active and easy to navigate 👍

Belty6979 5979 Thanks for putting this together! I was new to the Vercel Community, and this guide made it much easier to understand where everything is and how to get started. Looking forward to learning more and being part of the discussions. Appreciate the effort!

Amy Egan Glad to have you here!

Carter Turnbull 2nd this. It would be a great feature that would drastically improve the loop between non technical team members and UX/UI polishes.

Baptiste 3322 Please 🙏 would be awseome

Akshay Dodeja Give how Vercel Toolbar was created pre AI -- this feature is basically useless without Agents being able to access the data. Does anyone at Vercel engineering use vercel toolbar comments?

Amy Egan I wasn't able to repro, but maybe I'm looking in the wrong place. Can you show me where you see standard tier pricing for these requests? AI Gateway supports OpenAI tiers through the `service_tier` param. Looks like you're using it right. https://vercel.com/docs/ai-gateway/capabilities/service-tiers

Mo091336 4719 Hi Amy, Fortunately, the issue with the GPT-5.5 Flex and Priority tiers was resolved just a few days after I wrote my post, before you had a chance to reply. It appears that Vercel simply had not updated its billing system to reflect the new pricing structure, since the provider was returning responses with the correct service tier. However, the same issue reappeared with the release of the GPT-5.6 series, apparently for exactly the same reason. This time, there also seems to be an issue with the updated caching system. OpenAI has introduced both implicit and explicit prompt caching, but Vercel still appears to be using the previous configuration. As a result, implicit caching is being applied by default, with input charged at 1.25 times the standard rate and cache breakpoints not being respected. Could you please take a look at this? P.S. The same behavior occurs with both the Chat Completions API and the Responses API.

Mo091336 4719 Update: Now It works. Thank you Amy.

Anshuman Bhardwaj Hi @wajidkhann7788-6932, welcome to the Vercel Community! I think technically Vercel is capable of hosting any large application. You can see many of the largest companies in the world using Vercel on our https://vercel.com/customers page. SEO depends on the framework you choose and the content itself. Vercel can help in this area by serving the content fast. I see no issues in using Vercel. > This is not a legal advice about the content or legality of the app itself.

Ahmad Sajid Thanks for the insight, Vercel will be a great fit for what I’m building!

Theostafford7 3647 This looks like a Vercel SSL provisioning issue rather than a DNS problem. Since HTTP is routing correctly but HTTPS fails during the handshake, Vercel support should check whether certificate issuance or domain verification is stuck on their side.

Ryu Hi Den, TLS 1.3 and HTTP/3 are related, but they are not the same thing. TLS 1.3 means the HTTPS handshake can use the newer TLS version. HTTP/3 specifically means the site is being served over QUIC/UDP instead of HTTP/1.1 or HTTP/2 over TCP. A quick way to test what your site is actually negotiating is: ``` curl -I --http2 https://yourdomain.com curl -I --http3 https://yourdomain.com ``` or in Chrome DevTools: ``` Network tab → right click columns → enable Protocol ``` If you see `h2`, that is HTTP/2. If HTTP/3 is active, you’d usually see `h3`. From the current Vercel docs I can find, Vercel explicitly documents TLS 1.2 / TLS 1.3 support, but I don’t see HTTP/3 listed as a supported CDN protocol yet: https://vercel.com/docs/cdn-security/encryption So I would not assume HTTP/3 is available just because TLS 1.3 is enabled. This looks more like a platform feature request / roadmap question than something you can enable per project with `next.config.js` or DNS settings.

Den.D Hi @ryux1 , thanks fro the reply. You hit the nail on the head. As I mentioned in my post, even though the two things aren't directly connected, I don't understand the logic of supporting only TLS 1.3 and not HTTP/3. This is especially true when even the cheapest shared hosting providers ($10–$20/month) offer HTTP/3 by default. Of course, there's no personal setting for that, being a platform lack... My post was just a rant, hoping that someone in Vercel reads it and acts to solve the gap

Ryu Hi Den, Yep, fair point. I mostly replied to clarify the TLS 1.3 vs HTTP/3 difference for anyone else reading the thread. I agree there probably isn’t anything to configure at the project level right now. It would be nice to have HTTP/3 support on Vercel too, or at least a clear note in the docs about the current status. Thanks for raising it.

Wjh6548 4228 yeah, so am i. So I couldnt finished my rest assignment. how am i supposed to do??

Ryu Hi Bradley, `Failed to prepare VM` sounds like v0 is failing before it can fully start the sandbox/preview environment, so I’d first separate a broken project from a broader v0/account/platform issue. I’d test these in order: ``` 1. Create a brand-new blank v0 chat/project 2. Try the same action in an incognito/private window 3. Try a different browser with extensions disabled 4. Check whether it happens only on one project or every project 5. Check the browser DevTools console/network tab for the first failed request ``` If it happens only on one project, the project may have something that prevents the sandbox from starting, like an install script that hangs, a dependency conflict, a very large generated file, or a command that never exits. If it happens on every new blank project too, then it is probably not your app code. In that case the useful public details to share are: ``` Exact timestamp + timezone: Browser/device: Does it fail on a brand-new blank project: yes/no Does it fail in incognito/private mode: yes/no Any console/network error text, with tokens/cookies removed: Whether https://www.vercel-status.com/ showed any incident at that time: ``` I would avoid posting private project links, tokens, cookies, or account IDs publicly, but the timestamp and redacted console/network error would make this much easier to triage.

Ryu Hi Operendia, Your evidence points away from the app upload code and toward the public Blob serving path. If the Blob API can list the object with `HTTP 200`, but the direct public URL returns: ``` 503 SERVICE_UNAVAILABLE X-Vercel-Error: SERVICE_UNAVAILABLE Body: The deployment is currently unavailable ``` then this is not a normal “file missing” case. A missing blob would be a different kind of failure, and your URL format also matches the public Blob URL format Vercel documents: https://vercel.com/docs/vercel-blob/public-storage I’d do one isolation test before escalating it as store-specific: ``` 1. Create a brand-new public Blob store in the same team 2. Upload one tiny test image/text file with access: "public" 3. Open the returned public URL directly with curl -I ``` If the new store works but `pywla5knjbnow7ji.public.blob.vercel-storage.com` still returns 503 for every object, that strongly suggests the old store’s public serving layer is stuck or unavailable. If the new store also returns 503, then I’d check whether the team/project is paused, blocked, or over a Blob usage limit, because that would affect serving even if the token can still perform some API operations. The useful escalation summary would be: ``` Store ID: pywla5knjbnow7ji Blob API list/upload: 200 Public blob URL: 503 SERVICE_UNAVAILABLE Affects all blobs, old and new Tested from multiple networks Latest X-Vercel-Id: Latest timestamp: New public Blob store test result: Team/project paused or usage-limit banner: yes/no ``` Do not post the `BLOB_READ_WRITE_TOKEN`, but the public blob URL, `X-Vercel-Id`, and timestamp are safe and useful for Vercel to correlate.

Ryu Hi Abdelaziz, Since it works from personal devices and mobile but fails only on corporate devices with `DNS_PROBE_FINISHED_NXDOMAIN`, I’d first treat this as a DNS resolution issue on the enterprise network, not a Vercel deployment problem. Ask someone on the affected corporate network to compare the corporate resolver with a public resolver: ``` nslookup yourdomain.com nslookup yourdomain.com 1.1.1.1 nslookup yourdomain.com 8.8.8.8 ``` or: ``` dig yourdomain.com dig yourdomain.com @1.1.1.1 dig yourdomain.com @8.8.8.8 ``` What the results mean: ``` Corporate resolver returns NXDOMAIN, public resolvers return records → likely corporate DNS filter, DNS cache, split-horizon DNS, or security policy All resolvers return NXDOMAIN → likely domain DNS/delegation is not configured correctly Only www or only apex fails → check that both yourdomain.com and www.yourdomain.com are configured in Vercel/DNS ``` If only the enterprise DNS returns NXDOMAIN, the fix usually has to come from their IT/network team. They may need to clear DNS cache, allowlist the domain, or check whether a security product is blocking newly registered/uncategorized domains. I would also ask them to test the exact domain you expect users to open, for example: ``` https://example.com https://www.example.com ``` because apex and `www` can resolve differently.

Ryu Hi Pjmathur0291, For an HTTP-01 validation failure, I’d check the DNS and certificate-blocking records before assuming it needs a manual certificate flush. From a terminal, I’d run: ``` dig A raminstitute.in +short dig AAAA raminstitute.in +short dig CNAME www.raminstitute.in +short dig CAA raminstitute.in +short dig TXT _acme-challenge.raminstitute.in +short curl -I http://raminstitute.in/.well-known/acme-challenge/test ``` A few things to look for: ``` - The apex A record should match the value Vercel shows for the domain - The www CNAME should match the value Vercel shows, if you added www - AAAA should not point a third-party custom domain to Vercel - If CAA records exist, they need to allow letsencrypt.org - A stale _acme-challenge TXT record from an old provider can block issuance - /.well-known should not be redirected or rewritten by your app ``` Vercel’s domain troubleshooting guide covers those SSL checks here: https://vercel.com/docs/domains/troubleshooting If all of those are clean and the dashboard still shows HTTP-01 failure, the useful details to share are the exact domain status from Vercel, the `dig` output above, and the timestamp when you retried certificate generation.

Amy Egan After the high usage falls off from the rolling 30 day window, Vercel Agent can help you unpause the team https://community.vercel.com/t/my-account-is-still-paused-even-after-low-usage/43354/2