Proxyca team · 27 Oct, 2026 · Updated 31 Oct, 2026
A proxy that is not working may have a connection problem, invalid credentials or a restriction at the destination. Those failures need different fixes. Start with the exact error and a simple connection test before replacing the IP or changing several settings at once.
A website blocking a request also does not prove that your proxy account is broken. Separate the connection to the proxy, its connection to the destination and the destination's decision to accept your request.
Record the error, application, time and destination. Then check whether another ordinary destination works through the same proxy. This helps distinguish a general connection failure from a restriction affecting one service.
| What you see | What it may indicate | First check |
|---|---|---|
| Connection refused | Nothing accepts the connection at that address and port, or a network device rejects it | Confirm host, port and endpoint status |
| Connection timeout | The connection or response did not complete in time | Check connectivity, firewall rules and service status |
| HTTP 407 | The HTTP proxy requires authentication | Check credentials and supported authentication |
| SOCKS authentication failure | Proxy login or method negotiation failed | Verify protocol, login and endpoint requirements |
| HTTP 403 | The responding server refuses the request | Identify who returned it and read the response |
| HTTP 429 | The responding service applies a request limit | Respect Retry-After if provided and reduce request rate |
| DNS error | A hostname could not be resolved | Identify whether resolution happens locally or at the proxy |
These are starting points, not automatic diagnoses. An HTTP status can come from a proxy, gateway or destination. Preserve the response message and relevant headers so support can identify which component generated it.
First compare your settings with the current connection details in your account. Check for an old port, a copied space or a protocol mismatch. An HTTP proxy and a SOCKS5 proxy can use similar-looking host and port fields while requiring different handshakes.
Make sure your plan is active and any required traffic allowance remains available. If access uses an IP allowlist, check the public address of the network you are currently using. Moving between office Wi-Fi, mobile data and a VPN can change that address.
For a command-line test, curl can ask a SOCKS5 proxy to resolve the destination hostname. Replace the example host, port and username with your own values. Supplying the username without a password lets curl prompt for the password instead of placing it directly in the command history:
curl --proxy socks5h://proxy.example:1080 \
--proxy-user 'YOUR_USERNAME' \
--connect-timeout 10 --max-time 20 \
--output /dev/null --write-out 'HTTP %{http_code}\n' \
https://example.com/
This is a connectivity test, not an IP-address check or a benchmark. proxy.example is a placeholder and will not work until replaced. If your endpoint uses another protocol or authentication method, use its documented settings instead. The curl manual documents these options and their behaviour.
For a fuller verification sequence, use our guide to checking a proxy. Do not disable certificate verification to make an unexplained error disappear; inspect the hostname, system clock and certificate configuration instead.
An HTTP 407 response means the HTTP proxy requires authentication. It is different from a destination login failure. SOCKS proxies use their own negotiation and reply formats, so a SOCKS authentication failure may appear as a client error rather than an HTTP status.
Check the username, password and any documented username parameters. A generated residential connection may include options in its username; removing or mistyping them can change or invalidate the request. Copy a fresh connection from the account rather than reconstructing a complex string from memory.
If the application has separate credential fields, use them. If it expects a URL, characters with special meaning in URLs may need encoding. Follow the application's format and avoid posting complete connection strings in screenshots or public support threads.
Repeatedly submitting the same rejected credentials does not help diagnose the problem. Correct the configuration or ask the provider to verify the account and endpoint.
A 403 response indicates that the responding server understood the request but refuses to fulfil it. It does not, by itself, reveal whether the cause is an account permission, network policy or another access rule. Check the response body and the destination's documentation before drawing conclusions about IP reputation.
A 429 response indicates too many requests under the responding service's rules. RFC 6585 defines this status and allows a Retry-After header. Pause when instructed, reduce concurrency and avoid an immediate retry loop.
When no retry interval is supplied, use a limited retry policy with increasing delays and follow the service's published limits. Rotating IPs is not a substitute for respecting those limits. It may also disrupt an authenticated session or make diagnosis harder.
An account suspension is a separate issue. Review the platform's stated reason and use its support or appeal process. A replacement proxy does not restore account permissions or guarantee that subsequent activity will be accepted.
DNS resolution can happen on your device or at the proxy, depending on the protocol and client. In curl, socks5h:// requests proxy-side resolution, while socks5:// resolves the destination locally. A different application may express that choice differently.
If an address works but the hostname fails, inspect DNS settings and the exact hostname. Do not assume every DNS error is an IP block. Also distinguish the proxy's hostname, which your client must locate, from the destination hostname being requested through it.
Our SOCKS4 and SOCKS5 comparison explains the protocol differences. For sessions that change exit addresses, see how rotating proxies work.
Location databases can disagree, and a website may use account settings or other signals in addition to an IP address. Record the observed location and the service reporting it. A mismatch is useful evidence for support, but it is not a diagnosis on its own.
Give support enough detail to repeat the failure without exposing credentials. Include the endpoint host and port, protocol, application version, error message, timestamp with timezone and the destination you were testing.
Explain whether the same configuration works for another destination and whether the problem started after a change. For an HTTP error, include the status and relevant response headers. Redact passwords, session cookies, tokens and full authorization headers before sharing logs.
Ask for a specific check: endpoint availability, account authentication, traffic allowance or support for the required protocol feature. A focused report is easier to investigate than “the proxy is blocked.”
First determine whether the failure is network-related, an authentication problem or a destination restriction. A new IP will not correct a wrong port or an expired password. Replace an endpoint when the evidence and provider guidance support that action.
No. It confirms only the tested request. Another application can have different settings or unsupported traffic types. Verify the actual workflow after the basic test succeeds.
No. Websites can restrict requests for many reasons, including account permissions and request limits. Choose an appropriate service for a permitted task, and treat access decisions separately from connection quality.
Need help checking your Proxyca connection? Open Proxyca in Telegram and contact support with the diagnostic details above, with passwords removed.