A timeout in Shadowrocket doesn’t necessarily mean the server is down. This guide is for users who already have their own provider, nodes, or subscription. First, distinguish a failed test from a failed connection. Then narrow down the cause by testing the same node on another network and another node on the same network. Next, check protocol settings, system time, and routing rules. Change one thing at a time, and keep the original configuration so you can retest or revert.
1. First, tell a latency test timeout from a website loading failure
A latency test reflects only the specific test method, destination, and network used. A TCP connection test can show whether a particular TCP port is reachable, but it cannot by itself confirm that the password is correct, the TLS handshake succeeded, or the target website is available. A request sent through a proxy is closer to the actual browsing experience, but results can still depend on the test destination. Record the test method so you can compare results consistently.
Bottom line: reproduce the issue before making changes
If the latency test times out but you can access the same destination through that node, check the test method and destination before rebuilding the node. If browsing also fails, continue with network and protocol troubleshooting.
In Home, note the selected node, Global Routing mode, connection toggle status, and the time of the error. If you use a policy group, check which node is actually selected—not just the group name. Keep the destination website the same during testing so differences between websites don’t look like changes in node performance.
2. Use cross-tests to narrow down the cause
Turn off the Shadowrocket connection and confirm that your current network can access the internet. If the Wi-Fi network requires browser-based sign-in, complete that first. Then reconnect and test in the order below. Switching between Wi-Fi and cellular may incur data charges. On desktop, compare results using another network available to you.
-
Keep the test target the same
In Home, select an existing node and note the latency test method. Use the same method for each repeat test, and open the same target website separately to see whether it loads.
-
Switch networks
Keep the node and protocol settings unchanged. Switch from Wi-Fi to cellular, or to another available network, and repeat the same test. Don’t update your subscription during this test.
-
Try another existing node
Return to the original network and select another node you already have in Home. If you use a policy group, confirm that the group’s actual selection changes too, then test the same destination.
-
Note what changed
Be specific: “Node A fails on Wi-Fi but works on cellular” or “Only Node A fails on the same network.” Keep the exact error message instead of describing every failure as “doesn’t work.”
- The same node fails on only one network: Check that network’s DNS resolution, port reachability, and UDP support first. One test alone isn’t enough to conclude that the server is down.
- Only one node fails on the same network: Check that node’s settings, account status, and whether the provider is maintaining the server.
- Different nodes fail on every network: Check your device’s time, connection permissions, current routing mode, and any other active VPN connection. Multiple nodes may still be affected by the same issue if they share a server.
3. Check the address, port, and protocol settings
In Home, open the selected node’s edit screen and check each field against the original details from your provider. Address is usually a hostname or IP address; don’t paste a full webpage URL into the address field. Port must match the port the server is actually listening on. Port 443 is common for HTTPS, but that doesn’t mean every node should use 443.
| Check | Fields that must match | Common mistakes |
|---|---|---|
| Basic connection | Protocol type, Address, and Port | A trailing space in the address, a mistyped port, or the wrong protocol type. |
| Shadowsocks | Password and encryption method; check plugin settings too, if applicable | Checking the password but overlooking the encryption method or plugin settings. |
| VMess / VLESS | User ID, transport method, and related TLS settings | Mixing up settings for different protocols or missing transport fields such as Host or Path. |
| Trojan | Password, server address, and TLS settings | Changing the connection address and accidentally changing the server name required for the certificate. |
| Hysteria2 | Authentication details, port, TLS, and any other settings required by the server | Overlooking whether the current network allows the UDP traffic the protocol requires. |
| WireGuard | Local private key, peer public key, Endpoint, and tunnel address | Entering the public key and private key in the wrong fields, or using an outdated endpoint configuration. |
The TLS Server Name is used for certificate and handshake checks; it doesn’t necessarily match the IP address used to connect. The WebSocket Host and Path are also required settings, not optional decoration. If the server requires a specific path, a missing leading slash or incorrect capitalization can cause the connection to fail. Check these fields only if they’re enabled in the original configuration. Don’t add TLS or WebSocket settings to every protocol.
4. Distinguish connection, authentication, and certificate errors
The examples below are error messages that may appear during network failures. Exact wording varies by protocol, platform, and point of failure, and may differ from these examples. Keep the complete error shown in the app and use it to identify the type of problem.
Error: Connection timed out
Cause: A connection stage didn’t complete in time. Possible causes include the network path, an unresponsive server, or a stalled protocol handshake.
Fix: First, test the same node on another network. Then check the address, port, and relevant transport requirements. Don’t assume a timeout means the password is wrong.
Error: Connection refused
Cause: The connection was actively refused. Common causes include no service listening on the target port or access being explicitly denied.
Fix: Check that the port is still valid, and ask your provider to confirm that the server is listening. Repeatedly changing local DNS settings usually won’t fix a refused connection.
Error: certificate has expired
Cause: The certificate failed its validity check. The server certificate may have expired, or the device date may be incorrect.
Fix: Correct the system time and test again. If the time is right and the same error persists, ask your provider to check the certificate chain and expiration date.
If the error points to an authentication failure, check the password, user ID, and account status first. If it points to a DNS resolution failure, confirm that the node’s hostname is spelled correctly and that the current network can resolve it. Don’t change DNS, the port, the password, and routing mode all at once. Otherwise, even if the connection starts working, you won’t know which change fixed it.
5. Check system time and automatic time sync
An incorrect device date can affect certificate validity checks and authentication that relies on time. A different displayed time zone doesn’t necessarily mean the actual time is wrong. Check the date and current time; you don’t need to force every device to use the same time zone.
- iPhone / iPad: Go to “Settings → General → Date & Time” and check “Set Automatically” and the displayed time. If device management restricts this option, check the management policy first.
- Mac: Go to “System Settings → General → Date & Time” and check the option to set the date and time automatically.
- Windows: Go to “Settings → Time & language → Date & time.” Check that automatic time is enabled and sync now if the option is available.
- Android / Linux: Open the system date and time settings and confirm that network time sync is available. The exact menu depends on your device’s operating system.
After correcting the time, disconnect and reconnect Shadowrocket, then repeat the original test. If the device time drifts again after a restart, fix the system time-sync issue first. Changing the node password repeatedly won’t fix the device clock.
6. If connected but websites won’t load, check routing and On Demand
Turning on the connection only means the connection state has changed; it doesn’t mean every request goes through the selected proxy node. Actual traffic also depends on Global Routing mode, rules in Config, and policy group selection. For comparison, you can temporarily switch to Proxy mode, then restore your original settings when you’re done.
- Config mode: Traffic is handled according to the active configuration. Open Config and confirm which file is in use, then check the relevant domain and IP rules.
- Proxy mode: Use this to test the proxy route. If the same destination works in this mode but fails in Config mode, focus on rule matching and policy selection.
- Direct mode: Traffic isn’t normally routed through a proxy node in this mode. A website loading in Direct mode doesn’t prove that the selected node works.
- Scene mode: Behavior depends on the configured scene. Avoid switching scenes during comparison tests, and note the active mode before and after each test.
Rules are matched in order. Check the relative order of DOMAIN-SUFFIX, GEOIP, IP-CIDR, and FINAL rules. For example, if a destination matches DIRECT first, selecting a node won’t automatically route traffic through it. The example below shows rule syntax; it isn’t a general-purpose fix for timeouts.
[Rule]
DOMAIN-SUFFIX,example.com,DIRECT
IP-CIDR,192.0.2.0/24,DIRECT
FINAL,PROXY
In the example, PROXY must refer to a valid policy in your configuration; don’t overwrite your existing file. If the connection turns back on automatically after you switch it off, go to Settings → On Demand and check whether a relevant condition is enabled. For repeatable results, note the original conditions, temporarily turn them off for testing, then restore them. On Demand controls when connections start; it won’t fix incorrect protocol settings.
7. Update subscriptions and share useful troubleshooting details
Subscription updates and node connections are separate processes. An update can fail because the subscription URL is unreachable, and a successful update doesn’t mean every node in the list will connect. If you already have a subscription, note the node settings and what changed before and after the update, then use the subscription’s update action to retrieve the latest information from your provider. Don’t delete every subscription just to troubleshoot one node timeout.
Bottom line: report details about the stage that failed
If the issue persists after cross-testing, checking settings, and correcting the time, send your provider the network type, protocol type, time of the failure, and exact error message. Specify whether the failure occurred during DNS resolution, connection, or handshake; this is more useful than reporting only “timeout.”
- Include: Device operating system, Wi-Fi or cellular, test method, whether all nodes are affected, and the result of testing the same node on another network.
- Add: Current Global Routing mode, the policy group’s actual selection, certificate or authentication error type, and whether you recently updated the subscription.
- Hide: Full subscription URLs, tokens, passwords, private keys, and other authentication details. Check images and text for credentials before sharing your troubleshooting notes.
- Restore: When testing is complete, restore the original routing mode, On Demand conditions, and any temporarily changed settings so troubleshooting doesn’t affect everyday use.
This site doesn’t provide, sell, or recommend nodes or subscriptions. Contact your own provider about node authentication, account limits, server maintenance, and certificate renewals; getting the client and obtaining network service are separate matters. For more about configuration fields, see configuration fields. For the initial connection steps, see the setup guide.
Shadowrocket Install Options
The iPhone / iPad app is a one-time $2.99 purchase on the App Store. The Mac (Apple Silicon / Intel), Windows, Android, and Linux versions are free. For iPhone / iPad system requirements, see the App Store listing. Purchasing the app doesn’t include a network service plan.
Visit the download page for platform-specific install options. If you already have a configuration, continue to the first-connection and import steps.