Errors
Every HTTP status and SOCKS5 reply code the gateway returns, what each one means, and what to do about it.
The gateway reports problems as a normal HTTP status (for HTTP/HTTPS CONNECT) or a SOCKS5 reply code (for SOCKS5). Both protocols cover the same set of situations:
| HTTP | SOCKS5 reply | Meaning |
|---|---|---|
407 |
0x01 (auth failed) |
Unknown login, wrong password, paused credential or invalid username option. |
402 |
0x02 |
Allowance exhausted — no quota and no USD balance. |
429 |
0x02 |
Too many concurrent connections for your account. |
502 |
0x01 |
Upstream network rejected the connection (e.g. unsupported target). |
503 |
0x02 (maintenance) / auth status 0x01 (backend unavailable) |
Gateway in maintenance or temporarily unavailable. |
504 |
0x04 |
Upstream timeout while connecting to the target. |
407 — Proxy Authentication Required
Something about the credential is wrong: the login doesn't exist, the
password doesn't match, the credential is paused, or a -key-value
targeting option is malformed or out of order.
What to do: check the login and password under Proxies (rotate the password if you're not sure it's current), and verify the username suffix against Targeting — lowercase values, correct key order, no unrecognized keys.
curl -x "http://ofp_0123abcd:PASSWORD@proxy.oneforproxy.com:8080" https://api.ipify.org
402 — Payment Required
Your traffic package quota is used up and your USD balance can't cover pay-as-you-go billing either.
What to do: top up your balance or buy a traffic package on Billing — see Billing and packages and Pricing.
429 — Too Many Requests
You've hit the concurrent-connection limit for your account.
What to do: close idle connections, add connection pooling/reuse in your client, or lower how many requests you run in parallel.
502 — Bad Gateway
The upstream network refused the connection — often an unsupported or blocked target site, not a problem with your credential.
What to do: retry the request; if it keeps failing for the same
target, try different targeting (a different -country- or -asn-) or
confirm the target itself is reachable from a normal connection.
503 — Service Unavailable
The gateway is in maintenance or temporarily unavailable.
What to do: back off and retry after a short delay. This isn't something a credential or username change fixes.
504 — Gateway Timeout
The gateway couldn't connect to the target in time.
What to do: retry, and if it's persistent for one target, try a
different -country-/-asn- or a longer client-side timeout.
General advice
Treat 429, 503 and 504 as retryable with backoff, and 407/402 as
configuration problems to fix before retrying — hammering the gateway with
the same bad credential or exhausted balance just produces more 407s or
402s. Client-side examples that catch these are in
curl, Python and Node.js. The canonical version of
this table lives in the Docs.