Shadowrocket Connection Timeout Troubleshooting: Latency Tests, Protocol Settings & System Time

Troubleshoot Shadowrocket timeouts in order: test latency, verify protocol and encryption settings, then check your device’s clock.

At a glance

This guide is for users who have imported their own service details into Shadowrocket but are experiencing node timeouts or connection failures. Work through the checks in order—basic network access, latency tests, protocol fields, system time, and traffic routing—to distinguish an unreachable server, a failed handshake, and requests that aren’t using the selected node. Change one setting at a time and keep the before-and-after test results.

First, distinguish a timed-out test from a failed connection

The latency result shown in Home, the system VPN status, and whether a webpage loads are three separate signals. A connected VPN usually means only that the device’s local tunnel is established; it doesn’t confirm that the remote server has authenticated successfully. A successful latency test also doesn’t guarantee that every website or type of protocol traffic will work.

Think of the connection process as a chain of steps. A failure earlier in the chain can make later checks misleading. For example, changing the encryption settings won’t fix a server address that hasn’t resolved.

DNS resolution Transport connection Protocol handshake Authentication Destination access

Step 1: Keep the network fixed and repeat the latency test

First, confirm that the device can open a known working webpage. If the Wi-Fi network requires sign-in, complete its authentication first. The steps below refer to the Home and Settings interface labels on iPhone and iPad. Desktop layouts may differ, but record the selected node, test method, and network for each test.

  1. Record the current setup

    In Home, note the selected node and Global Routing mode, and save the original configuration. Don’t refresh the subscription, change DNS, and switch protocols all at once—you won’t know which change affected the result.

  2. Keep test settings constant

    Go to Settings → Connectivity Test and check the current test method and any related destination settings. If the method uses a test URL, make sure the URL itself is reachable. Keep the method and destination unchanged while troubleshooting.

  3. Repeat the test

    Return to Home and run Connectivity Test on the same node three times in a row. Record each result or timeout. One successful test doesn’t prove the connection is stable, and one failure doesn’t mean the server is down.

  4. Switch networks

    If possible, switch from Wi-Fi to cellular and run the same test on the same node three times. Change only the network connection; don’t alter the address, port, protocol, or test destination.

  5. Test actual access

    Select the node, connect, and try to access a known working destination. Record the exact time and what happened. If automatic connection interferes, note the Settings → On Demand status, temporarily turn it off while troubleshooting, then restore it afterward.

The test method determines what the result means. A TCP connection test only shows that a connection can be established to the relevant TCP port. An HTTP request sent through the proxy also involves a protocol handshake and access to the destination. No ICMP response doesn’t necessarily mean the proxy is unavailable; the server may not respond to this type of probe. For protocols that rely on UDP, a TCP port test can’t replace a full connection test.

Observed resultCheck firstNext step
The same node fails on Wi-Fi but works on cellularWi-Fi authentication, DNS, gateway, or UDP restrictionsKeep the node settings unchanged and check the current network
The same single node fails on both networksThe node’s address, port, and server statusConfirm the connection details with your service provider
TCP test succeeds, but the actual request failsAuthentication, TLS, transport settings, or destination accessCheck the fields for the specific protocol
Multiple nodes fail only with the same test URLWhether the test destination is reachableConfirm the destination, then test access separately

Takeaway: Keep the results from testing across networks

If the same settings work on cellular but not Wi-Fi, check the Wi-Fi connection path before rebuilding the node. This comparison narrows down the cause more effectively than a single latency reading.

Step 2: Check the address, protocol, and authentication fields

In Home, open the selected node’s editing screen and compare each field with the original information from your service provider. Check Type, server address, and port first, then review authentication and transport settings. The server address field usually takes a hostname or IP address. Don’t enter a full subscription URL as the server address or add a path unless the field requires one.

For example, if the service details specify port 443, the client and server must use the same port. But port 443 doesn’t prove that the node uses TLS, and it’s not a reason to change the protocol to Trojan. Port, protocol, and security settings are separate fields; use the original configuration as your reference.

ProtocolFields to checkCommon misconception
ShadowsocksAddress, port, encryption method, password, and any plugin parameters required by the configurationThe password may be correct, but a different encryption method can still prevent communication
VMessUUID, encryption settings, transport method, TLS, and any other fields required by the serviceCopying the UUID but missing the transport path or Host
VLESSUUID, transport method, TLS or required security settings, and flowReusing another node’s flow or security settings without checking
TrojanPassword, TLS, Server Name, and transport settingsConfusing the connection address with the name required for certificate verification
Hysteria2Authentication, TLS, Server Name, and any service-specific obfuscation settingsOverlooking that it uses QUIC and requires UDP
WireGuardLocal private key, peer public key, Endpoint, interface addresses, and allowed routesChecking only the Endpoint address while overlooking key pairing and tunnel routing

When using transports such as WebSocket, paths are usually case-sensitive, and the leading slash must be preserved. Server Name identifies and verifies the server in TLS connections; don’t clear it just because the connection address is an IP. Don’t disable certificate verification as a permanent fix. First check the name, certificate validity period, and system time.

Step 3: Check system time and TLS errors

TLS certificate verification uses the device’s current time. If the clock is significantly ahead or behind, a valid certificate may appear not yet valid or expired. Some authentication methods are also sensitive to clock skew. If the protocol fields look right but the handshake keeps failing, turn on automatic time setting and repeat the same test.

The following are examples of messages that may appear in network or TLS logs. Exact prefixes and wording vary by system and connection library. Check the full context rather than changing settings on every node based on a single word.

Error: i/o timeout

Cause and fix: A network read or write operation didn’t finish in time. This message alone can’t tell you whether the port or destination website is at fault. Check the address and stage in the surrounding log entries, then compare results across networks to identify the relevant connection path.

Error: connection refused

Cause and fix: The destination host or an intermediate device actively refused the connection, possibly because no service is listening on that port. Check the address and port, and ask your service provider to confirm that the port is listening. Repeatedly changing the password won’t fix this.

Error: certificate has expired or is not yet valid

Cause and fix: The certificate’s validity period doesn’t match the current time. Sync the system clock first. If the error persists with the correct time, ask the server administrator to check whether the certificate needs renewal.

Error: certificate verify failed

Cause and fix: Certificate verification failed, possibly because of a name mismatch, an incomplete certificate chain, or a trust issue. Check that Server Name matches a name on the certificate and review the server’s certificate chain. Don’t bypass verification.

Step 4: Confirm that requests use the selected node

Node tests and everyday requests may take different paths. In Home, check Global Routing: Proxy uses the proxy mode, Direct connects directly, Config follows configuration rules, and Scene follows the logic set for that scene. Selecting a node in Home doesn’t guarantee that every request will use it.

For a temporary test, note the original mode, then briefly switch to Proxy and check the same destination. Restore the original setting afterward. If Proxy works but Config doesn’t, check the active configuration and matching rules in Config instead of continuing to change authentication fields that have already been verified.

[Rule]
DOMAIN-SUFFIX,example.com,DIRECT
IP-CIDR,192.168.0.0/16,DIRECT,no-resolve
GEOIP,CN,DIRECT
FINAL,PROXY

The following example illustrates rule matching only; don’t use it to replace your existing configuration. Rules are checked from top to bottom, and processing stops when one matches. DOMAIN-SUFFIX matches domain suffixes, IP-CIDR matches IP ranges, GEOIP uses IP geolocation, and FINAL handles requests that don’t match earlier rules. In this example, example.com and its subdomains match DIRECT first, so they won’t reach the final PROXY rule.

Common questions and troubleshooting details to share

Latency is only a few dozen milliseconds. Why won’t webpages load?

First, check the test method in Settings → Connectivity Test. If it only confirms a successful TCP connection, also check protocol authentication and TLS. Then return to Home, check Global Routing, and try the same destination directly. Record when the connection fails.

Hysteria2 works on cellular but always times out on Wi-Fi. Why?

Keep the node settings unchanged and repeat the test on both networks. Hysteria2 relies on UDP, which some Wi-Fi networks may restrict. Ask the network administrator whether UDP is reachable; a TCP port test isn’t a complete substitute.

It still fails after refreshing the subscription. Should I keep refreshing?

No need to refresh repeatedly. First check whether the retrieved address and port have changed, then test one node. Updating a subscription and establishing a proxy connection are separate processes. If the update request itself times out, check that the link is valid and that the current network can retrieve it.

What information should I send to my service provider?

Include the time and time zone of the failure, operating system, protocol, network type, test method, results from comparing networks, and relevant error text with credentials removed. Don’t send a full subscription link, password, UUID, private key, or QR code containing authentication details.

A useful report should reproduce one specific issue. For example: “Same node and test method: timed out three times in a row on Wi-Fi and succeeded three times on cellular; automatic time setting is on; the actual request failed during the handshake.” This is an example of how to report results, not data collected by this site. If an error appears only once in the logs, say whether you can reproduce it.

Download Shadowrocket