At a glance

This guide is for users who already have their own service details and are seeing timeout errors in Shadowrocket. First, check your installation option: iPhone and iPad users can get it on the App Store for a one-time $2.99 purchase; see the App Store listing for system requirements. The Mac (Apple silicon or Intel), Windows, Android, and Linux versions are free; choose your platform on the all-platform downloads page. The steps below narrow down the cause by checking network reachability, protocol settings, system time, and actual requests—instead of judging service availability by a single latency result.

1. Distinguish a latency-test timeout from a failed request

timeout means an operation did not finish within the allotted time. It does not confirm that the server is at fault. A timeout can occur during address resolution, transport connection setup, a protocol handshake, or a request to the destination website. Start by noting where the error appears; don’t immediately change DNS settings or reinstall the app.

A test result in Home is not the same as browsing in a web browser. TCP-based tests mainly check whether a transport connection can be established. HTTP-based tests also depend on the test URL, protocol handshake, and subsequent request. The exact coverage depends on the selected test method. A latency result in milliseconds does not confirm that authentication settings are correct or that every destination website is reachable.

Where you see itWhat it tells youWhat to check next
Home test shows timeoutThis test did not finish within the allotted timeNote the test method, then retry the same entry on another network
TCP test returns a time in millisecondsThe TCP connection attempt received a responseCheck authentication and TLS settings, then make an actual request
Connection is on, but webpages fail to loadAn enabled connection does not mean requests are succeedingCheck Global Routing and how the destination request is handled
Only subscription updates time outThe subscription URL could not be retrievedCheck the subscription URL and its access path separately

Record three consecutive tests on the same network, including the time, network type, test method, and exact error text. Three tests are simply a practical sample for comparison—not an app default or a measure of connection quality.

2. Set up a consistent, repeatable test

Before you begin, save your current configuration and note the selected entry in Home and the Global Routing mode. Change one thing per round and record the result. If you switch networks, update a subscription, and change protocol settings all at once, it will be difficult to tell what made a difference—even if you later undo the changes.

  1. Confirm that the network works

    Turn off the connection switch in Home and use a browser to confirm that your current Wi-Fi or mobile network has internet access. If the network requires sign-in through a captive portal, complete that step before returning to Shadowrocket.

  2. Note the selected entry

    In Home, record the selected entry’s name, Type, server address, and port. The name is only a label; check the actual fields on the entry’s edit screen, and don’t identify a protocol by its name alone.

  3. Run a latency test

    In Home, run the latency test available for the same entry and record the test method and result. Don’t compare results from different test methods directly.

  4. Switch networks

    Switch between Wi-Fi and mobile data without changing the entry or test method. If only one network fails, first check its sign-in requirements, access restrictions, and UDP reachability.

  5. Test an actual request

    Return to Home, select the entry you want to test, check Global Routing, and then turn on the connection. Retry the same test webpage and note any errors. Restore the original mode when you finish testing.

3. Check the address, port, and protocol settings

Check each field against the service details you already have; don’t copy settings from another profile. Enter the host name or IP address in the server field—not the full share link—and don’t append the port to the host name. Use the port specified by the server: if it specifies 443, enter 443. Using that port does not necessarily mean TLS is enabled.

TypeFields to checkCommon misconception
ShadowsocksAddress, port, encryption method, password, and any service-specific parametersA successful TCP connection does not confirm that the password and encryption method match
VMess / VLESSUUID, transport, TLS, and any applicable Host, Path, and SNI valuesChecking the UUID but missing transport settings
TrojanPassword, TLS, SNI, and the name covered by the certificateAssuming the correct port guarantees a successful TLS handshake
Hysteria2Authentication details, TLS, SNI, and any applicable obfuscation settingsUsing a TCP test to judge whether UDP/QUIC works
WireGuardKey, peer public key, Endpoint, interface address, and route rangesBlaming webpage DNS when a UDP Endpoint gets no response

Host, Path, and SNI serve different purposes and cannot replace one another. Keep the leading slash and capitalization in Path; SNI should match the service’s requirements. Temporarily changing the server hostname to an IP address may change both certificate name validation and the connection destination. So a response after switching to an IP is not, by itself, a permanent fix.

If the entry comes from a subscription, keep a copy of its current details before updating your subscription, then compare what changed. Manual edits may be overwritten by a later update. Avoid importing multiple entries with the same name and testing them at random; you may still have the old address selected. This guide does not provide, sell, or recommend nodes or subscriptions. All checks are based on service details users already have.

Bottom line: check authentication after confirming reachability

If a TCP test gets a response but an actual request fails, check protocol authentication, transport, and TLS settings next. With Hysteria2 or WireGuard, also check UDP reachability separately; a TCP result cannot rule out network restrictions.

4. Narrow down the cause by the exact error

The messages below may appear in diagnostic information. Capitalization and surrounding text can vary with the system and connection stage, so keep a copy of the full error exactly as shown. These messages can narrow down what to check, but no single word confirms the server’s status.

Error: timeout

Cause: The operation did not finish within the allotted time. The delay may occur during resolution, transport, or a subsequent request.

What to do: Note where the error appeared, then retry on another network using the same entry and test method. Avoid changing multiple settings at once.

Error: connection refused

Cause: The connection was explicitly refused, often because nothing is listening on the destination port or a device along the path actively rejected it.

What to do: Check the server address and port, then ask your service provider whether the corresponding service is running. Don’t try to fix it by randomly changing ports.

Error: TLS handshake timeout

Cause: The TLS handshake did not finish in time. The cause may involve the network path, server response, or TLS settings.

What to do: Check TLS, SNI, and transport settings, then retry on another network and verify the system clock. Don’t disable certificate verification.

If the error specifically points to name resolution, first determine whether it is the server hostname or the destination website hostname that cannot be resolved. The former may prevent the service connection from being established; the latter may occur after a connection is already established. Then check DNS-related settings in Settings and the active configuration in Config. Don’t assume every timeout is a DNS issue.

5. Check the system clock, then review Global Routing

System time is used to check certificate validity periods, and authentication mechanisms such as VMess may also rely on a time window. A significantly incorrect clock can cause authentication or TLS to fail even when the server address and port are correct. But a correct clock does not prove that other settings are right. Check the time when there is a specific reason to do so; it is not a guaranteed fix.

Check that the actual date and time are correct, not just the time zone name. After correcting them, disconnect and reconnect Shadowrocket, then repeat the same test. If the error persists, keep the original error text and don’t attribute every subsequent failure to the clock.

Next, return to Home and check Global Routing. Proxy, Direct, Config, and Scene determine how traffic is handled overall; they are not four latency-test methods. A webpage loading in Direct does not prove the selected proxy entry works. In Config, check the active rules to see where the request is routed.

ModeWhat to check
ProxyTo isolate rule-related issues, make sure the entry you’re testing is selected. Restore the original mode when testing is complete.
DirectPrimarily useful for checking the direct connection path; it does not prove the selected proxy entry works.
ConfigCheck the active configuration’s rules and their order to determine which policy handles the request.
SceneCheck the conditions for the active Scene and the configuration or policy it actually uses.

For example, DOMAIN-SUFFIX,example.com,DIRECT in Config sends matching domains through DIRECT, not the proxy entry under test. Rules are generally evaluated in order. GEOIP and IP-CIDR apply to destination IP conditions, while a final FINAL rule handles requests not matched earlier. Check which policy matches before deciding that an entry is faulty; a request that bypasses it is not evidence that it works.

6. Common questions and what to share with your provider

Why does the latency test return a value if webpages still time out?

First, check which test method Home used. A successful TCP connection does not verify full protocol authentication. Next, check Global Routing, the entry’s authentication details, and TLS settings. If only specific domains fail, check their resolution and which rules match.

Wi-Fi times out, but mobile data works. Should I import the entry again?

Don’t import it again yet. Keep the same entry and test method, and make sure Wi-Fi sign-in is complete. If you use Hysteria2 or WireGuard, check whether that network allows the required UDP traffic. Then share the comparison with your service provider.

Only subscription updates time out. Does that mean all the entries are down?

No. Retrieving a subscription and connecting through an entry are separate processes. First, test an existing entry in Home. Then ask your service provider whether the subscription URL is valid and what access requirements apply. Don’t delete all existing entries before trying an update, and don’t share URLs containing tokens publicly.

Can I fix a TLS timeout by turning off certificate verification?

Don’t use disabled certificate verification as a routine troubleshooting step. First check the system clock, SNI, the domain covered by the certificate, and the server configuration. If there is a certificate issue, the service provider should fix it. Turning off verification won’t fix a network connection that gets no response.

If you still can’t find the cause, share reproducible details with your service provider. “This error occurs repeatedly on this network with this test method” is more useful than “it won’t connect.” Don’t include full subscription URLs, passwords, UUIDs, or private keys in public screenshots or discussion forums.

Shadowrocket Downloads and Setup Guide

The iPhone / iPad version is a one-time $2.99 App Store purchase; the Mac / Windows / Android / Linux versions are free. Mac downloads are available for Apple silicon and Intel. See the App Store listing for iPhone / iPad system requirements. Choose your device on the downloads page, then follow the setup guide to get started.