Run the four checks in order
| Check | How | What it rules out |
|---|---|---|
| DNS | Resolve the domain from the shell; try a second resolver | Name resolution, which fails before any HTTP happens |
| Your proxy | Turn the proxy and TUN off, then load the site | Your own tunnel being the reason |
| Your network | Try mobile data or another connection | A local or carrier-level block |
| The site itself | Check the provider's announcement channel and whether other users report the same | An outage, which no local change will fix |
Each check is cheap. Doing them in order prevents the most common mistake: changing the client configuration to fix a website that is simply down.
Why your proxy may be the cause
If you reach the dashboard through the provider's own subscription, a broken tunnel also breaks the dashboard. That produces a genuinely confusing state — the service appears gone while you are looking at it through the service. Turning the proxy off, or using a different route temporarily, separates the two.
This is also why a provider's status page should ideally be reachable without their own network. If it is not, treat its absence as uninformative rather than as confirmation of an outage.
DNS is more common than people expect
- Resolution failure looks identical to a dead site in a browser.
- A resolver that is unreachable because of the tunnel returns no answer, not an error you can read.
- Split-horizon DNS on a corporate network can make one domain resolve differently from another device.
Resolve from the shell, with the proxy off, and compare two resolvers. If both fail while other domains work, the domain or the resolver path is the question — not the provider's servers.
If it is an outage
- Check the provider's own announcement channel, and note the time you first observed the failure.
- Do not act on instructions that arrive only through a direct message: a “new domain” or “new client” link is a standard impersonation pattern.
- Do not send payment to a different address because the site is down.
- Keep your existing subscription; an outage of the dashboard does not necessarily mean the nodes stopped.
If the site loads but you cannot log in
That is a different failure, and a better one: the service is reachable. Check the session, the password reset path, and whether your account is locked. Do not reset your subscription token as a first response — that invalidates a working link for a problem that has nothing to do with it.
What to record
- Time, and the exact browser or resolver error.
- Whether the site loads with the proxy off, and on another network.
- Whether other domains resolve and load at the same moment.
- Whether the provider's announcement channel says anything, and when.
- Whether the login page is reachable separately from the dashboard.
Scope
This page describes an access-check order. It names no provider and makes no claim about any service's availability. A dashboard failure and a node failure are separate events, and treating them as one is what turns a short outage into an unnecessary reconfiguration.