This guide is for cases where Shadowrocket is already configured but web pages load slowly, videos buffer, or connections time out intermittently. Keep test conditions consistent, then check whether the server is available, the route to it is stable, the protocol settings match, and your local network is unrestricted. Change one thing at a time and keep the original settings so you can share useful details with your service provider.
Tell connection failures, slow responses, and low throughput apart
“Slow” can describe several different problems. DNS resolution, connection setup, TLS negotiation, and data transfer happen at different stages. A long wait before a page starts loading doesn’t necessarily mean downloads will stay slow. Likewise, a successful Connectivity Test doesn’t guarantee stable large-file transfers. First identify which stage is slow, then choose what to check.
This flow helps identify where to troubleshoot; it doesn’t mean every request follows the exact same sequence. The actual DNS resolution point can vary when an IP rule matches or proxy-side DNS is used. Before testing, pause system updates, cloud backups, and other bandwidth-heavy tasks. Use the same device, destination, and test window throughout.
- Nothing loads: First check the default server and connection toggle, then look for timeouts, authentication errors, or TLS errors. Don’t start by changing bandwidth-related settings.
- The first visit is slow, but later visits are normal: Check whether this happens with every new domain. Start by investigating DNS resolution and connection setup.
- Pages load, but transfers stay slow: Record the speed during sustained transfers and check for route congestion, provider throttling, and local bandwidth usage.
- Only one website is slow: Check the rule matched by that request and the status of the destination service. Don’t assume all proxy connections are affected.
Layer 1: Check the default server and actual outbound route
Seeing a server in Home doesn’t mean it’s handling the current request. Check which server is selected by default, then review Global Routing. Also check for policy groups in your configuration: a rule may still use a different, fixed outbound route even after you change the default server.
-
Record the original settings
In Home, note the current default server and Global Routing option. If you’re using Config, note which profile is enabled there as well. Restore these settings after testing.
-
Check the selected server
In Home, select one of your existing servers and confirm the selection indicator appears next to it. Turn on the connection. If iOS asks to add a VPN configuration the first time you connect, follow the system prompt.
-
Set up a comparison
In Home → Global Routing, temporarily select Proxy and test the same destination. This comparison helps rule out differences in Config rules. Don’t change DNS or protocol settings at the same time.
-
Test connectivity
Return to Home and run Connectivity Test, then open the website or content that’s actually causing the problem. Record separately whether the test completes and whether the content loads reliably; these results aren’t interchangeable.
-
Restore and test again
Restore the original Global Routing option, then visit the same destination. If Proxy works but Config doesn’t, check the matched rule and policy next instead of continuing to switch servers.
In Global Routing, Proxy tests outbound connections through a proxy; Direct tests a direct connection; Config routes traffic according to its rules; and Scene determines behavior based on the configured scenario. If a website opens with Direct, that only confirms its direct connection works—it doesn’t show that the proxy server is working. Results under Scene also depend on the scenario settings.
- If another existing server from the same provider works while the original one keeps failing, ask the provider to check the original server’s status, account validity, and traffic allowance.
- If all servers fail, check shared factors first, such as an expired subscription, whether your local network is working, and whether another VPN configuration is active on your device.
- Get subscription links from your service provider. After updating a subscription, check which server is selected by default; server names or entries may have changed.
Layer 2: Compare one variable at a time to pinpoint route issues
A server that connects successfully doesn’t guarantee a stable route between your device and the server. Cross-network routing, evening congestion, and packet loss can all reduce transfer speeds. A short connection test mainly reflects the response to one connection; it can’t capture fluctuations during a sustained transfer.
On the same network, use two of your existing servers to access the same content. Observe each for about 60 seconds, repeat three times, and note when you tested. Don’t download multiple files at once. Browser caching can make a second visit faster, so use the same cache conditions for each retest to avoid mistaking a cache hit for an improved route.
| Comparison setup | What you observe | Next step |
|---|---|---|
| Same Wi-Fi and destination; change only the server | Only one server stays slow | Ask about that server’s load, speed limits, and route to the destination |
| Same server and destination; change only the network | Slow on Wi-Fi, normal on cellular | Check the router, wireless interference, and broadband route to the server |
| Same server and network; change only the time of day | Consistently slower during peak hours | Keep results from multiple time slots and ask your provider to investigate congestion |
| Same server and network; change only the destination | Only one destination is slow | Check the destination service, exit location, and rule matched by the request |
For example, “Server A is slow on evening Wi-Fi but works normally on cellular” describes a useful troubleshooting observation; it isn’t a measured result from this guide. It narrows the possibilities to local Wi-Fi, broadband access, or the broadband route to Server A. That observation alone can’t establish that the server is at fault.
Layer 3: Verify protocol settings instead of switching protocols at random
The protocol must match the server configuration. Changing a VMess entry to VLESS or turning on TLS won’t automatically improve speed. Some incorrect settings prevent connections altogether; others can cause repeated reconnects. Open the existing server’s edit screen and compare each setting with the original details from your provider.
| Protocol | Check first | Troubleshooting notes |
|---|---|---|
| Shadowsocks | Address, port, encryption method, password | The encryption method and password must match the server; don’t omit any additional plugin parameters |
| VMess / VLESS | UUID, transport, TLS, SNI, Host, Path | Fill in only the fields required for that entry; required fields vary by transport |
| Trojan | Password, port, TLS, SNI | Verify the domain covered by the certificate. Don’t bypass the problem by disabling certificate verification |
| Hysteria2 | Authentication, TLS, server port | Uses QUIC and UDP; check whether the local network restricts this traffic |
| WireGuard | Key, Endpoint, tunnel addresses, AllowedIPs | Check the handshake and routing; don’t reuse authentication fields from other protocols |
A port number alone isn’t enough to identify a protocol. Even if your provider specifies port 443, that doesn’t tell you which protocol or transport to select. Check copied values for leading or trailing spaces, preserve the capitalization of Path, and make sure SNI hasn’t been mistakenly set to the server’s IP address.
If an error points to TLS, check the system date and time first, then verify the domain and certificate. If only Hysteria2 or WireGuard fails on one network but works on the same device with another network, a restriction on the UDP route is worth investigating. Ask your provider for a working configuration instead of changing ports at random or guessing at settings.
Error: connection refused
Cause and fix: The destination is reachable, but the connection to the specified port was refused. Check the address and port, confirm the server is listening, and ask your provider whether the entry is still available.
Error: i/o timeout
Cause and fix: A network read or write didn’t complete within the time limit. Possible causes include packet loss along the route, filtering, or an unresponsive server. Keep the server unchanged, switch between Wi-Fi and cellular, then compare with another existing server.
Error: TLS handshake timeout
Cause and fix: TLS negotiation didn’t complete in time, possibly due to the route or server-side TLS handling. First check TLS, SNI, and the system time, then share the time of the failure and the relevant error excerpt with your provider.
These are network error excerpts; a full log may also include an address, port, and higher-level error prefix. Error text can help narrow down the cause, but shouldn’t be used alone to identify one definitive cause.
Layer 4: Check Wi-Fi, cellular data, and reconnection settings
Start by checking the local network outside the proxy. Temporarily disconnect and visit a website you normally access directly to confirm that basic internet access works. Public Wi-Fi may require authentication first; changing server settings won’t usually help until you sign in.
- Wi-Fi: Test again near the router to tell a weak signal from a broadband route issue. If your device supports it, compare 2.4 GHz and 5 GHz. 5 GHz isn’t necessarily more stable through walls.
- Cellular: Check your signal and data allowance, and see whether your carrier plan has been throttled. Don’t mistake a brief interruption while switching from Wi-Fi to cellular for a persistent problem.
- Other devices: Check whether devices on the same network are running cloud backups, game updates, or sustained uploads. When upload bandwidth is saturated, web requests and download acknowledgements can also be delayed.
- Other VPNs: Check the active configuration in your system’s VPN settings to make sure another VPN connection isn’t replacing it during testing.
- Automatic reconnection: Check the connection conditions under Settings → On Demand. If you need to disconnect manually for a comparison, note the original settings, temporarily disable On Demand, and restore it when you’re done.
If small pages load but larger transfers stall, note whether this happens only with a particular protocol or network. Packet loss or MTU issues may be involved, but “video buffering” alone doesn’t prove an MTU problem. Start with the settings provided by your service provider. If you need to adjust anything, change one setting at a time and keep its original value.
If only Config is slow, check the rules and DNS path
If Proxy works but Config is noticeably slower, focus on the enabled profile in Config. Rules are generally evaluated from top to bottom, and the first matching rule determines the policy. Check whether an earlier rule sends the destination to Direct, another proxy policy, or a policy group whose name has changed.
[Rule]
DOMAIN-SUFFIX,example.com,PROXY
IP-CIDR,192.0.2.0/24,DIRECT,no-resolve
GEOIP,CN,DIRECT
FINAL,PROXY
This configuration illustrates rule order; don’t use it to replace your existing file. example.com and 192.0.2.0/24 are example destinations. DOMAIN-SUFFIX matches domain suffixes; IP-CIDR matches IP ranges, and no-resolve in the example avoids an additional DNS lookup for that rule; GEOIP matches IP geolocation; FINAL catches requests not matched earlier.
- Find the connection record for the affected destination and check which policy it actually uses before changing the relevant rule. Don’t infer the match based only on the destination’s geographic region.
- Confirm that custom policy names still exist. If a subscription entry was renamed, check any references to its old policy name.
- If every new domain is slow to load the first time, record your current DNS settings and compare DNS resolution separately. Changing DNS generally won’t fix slow transfers once a connection is established.
- Don’t replace DNS, import an entirely new rule set, and switch servers all at once. Retest the same destination after each change so you can interpret the results.
Common questions and what to include in a troubleshooting report
Why does video still buffer when Connectivity Test shows a low number?
The test reflects the response from its target using its particular test method; it doesn’t measure sustained video throughput. Keep the same server and record both the initial page response and transfer performance over about 60 seconds, then compare with another existing server.
It works on cellular. Do I need to change my configuration?
Keep the configuration for now. Compare Wi-Fi and cellular again using the same server, and check Wi-Fi authentication, router load, and the broadband route. Correct protocol settings only if they don’t match the information from your provider.
It got slow right after I updated my subscription. Should I add it again?
First check the default selection in Home, then review Global Routing and policy group references. An update may change entry names or the available server list; there’s no need to delete the entire configuration right away. If the issue continues, ask your provider to check the entry’s status.
It’s only slow at night. Is my phone underpowered?
Compare different times of day first. Keep the device, network, and server unchanged, and repeat the test both off-peak and in the evening. If results differ consistently, compare with another network and share the findings with your provider. Don’t blame the device based on a single observation.
A reproducible report is more useful for troubleshooting than saying “everything is slow.” Use the checklist below to organize the details, then send them through your service provider’s support channel.
- When and where: Local time and time zone, Wi-Fi or cellular, device operating system, and how long the issue lasts.
- Connection settings: Protocol, Global Routing option, whether On Demand is enabled, and any changes made before the issue started.
- Steps to reproduce: Which redacted server entry you selected, the type of content you accessed, and how long it took for the slowdown or error to appear.
- Comparison results: Whether another existing server, another network, or another time of day works; and whether the issue happens every time.
- Error excerpt: Include only what’s needed to investigate. Redact full subscription links, tokens, UUIDs, passwords, private keys, and personal addresses.
When you’re done troubleshooting, restore the original Global Routing option, On Demand setting, and any other settings you changed. This site doesn’t provide, sell, or recommend servers or subscriptions. Contact your service provider about server load, account allowances, or server-side settings.