Data shows traffic recorded by Shadowrocket; it is not your provider’s plan balance. The server view shows which proxy outlet traffic used, while the app view shows traffic attributed to apps the system can identify. To verify a rule, also check the connection log for the destination, matched rule, and final policy. This guide is for users with an imported subscription and a working connection who want to understand differences in the statistics or troubleshoot routing results.
1. Distinguish Data, system usage, and provider billing
Shadowrocket’s Data statistics have a specific scope. Readings depend on whether app requests pass through its processing chain, when tracking starts, and whether the counters have been reset. A figure in Data does not directly tell you how much cellular data the device used this month or how much data remains in your subscription.
Start with the direction: Upload is data sent from your device; Download is data received by your device. Browsing a webpage also uses upload, since request headers, query parameters, and connection exchanges must be sent. Downloads may account for most of the traffic, but uploads should not be expected to be zero. Likewise, uploading a file involves receiving responses from the server.
Use the table below to compare these three measurement sources. Before comparing figures, align the device, network type, and time period. If system usage covers cellular data while Data also includes Wi-Fi traffic, the totals will differ even when everything is working as expected.
| Where to check | What it can answer | What you cannot conclude directly |
|---|---|---|
| Shadowrocket Data | How recorded traffic is distributed and how much it increased during a test | Total network activity across the entire device |
| System network usage | How much of a specified network type the device or app used during a system-defined period | Whether a particular Shadowrocket rule matched |
| Provider account page | Usage and balance calculated under the provider’s billing rules | How much traffic each local app used |
2. By server and by app: two ways to group the same traffic
After checking the connection status on Home, open Data and switch between the server and app views. The entry points and available grouping options may vary by platform, so use the views available on your current screen. In the server view, focus on the selected server’s Upload and Download changes. In the app view, focus on records that the system and client can identify and attribute to an app.
By server
- Grouped by
- Traffic recorded for a proxy server
- Useful for
- Checking whether test traffic reached the expected proxy outlet
- How to read it
- Compare Upload and Download for the same server before and after a test
- Key limitation
- A proxy server’s figures do not represent all DIRECT traffic
An increase for a server only shows that traffic was recorded under that entry; it does not, by itself, prove that a rule for a particular webpage is correct.
By app
- Grouped by
- Connection traffic identified and attributed to an app
- Useful for
- Finding the main source of background sync, video playback, or file transfers
- How to read it
- Use one app at a time, then compare its change in the statistics
- Key limitation
- System capabilities and traffic attribution limits may prevent some records from being identified
If an app has no separate entry, do not assume that “unidentified” means “offline.”
These are two ways to view traffic, not separate sources to add together. A browser downloading a file through a proxy server may appear under both the browser and the server. Adding the two figures would count the same traffic twice. Conversely, differences in attribution coverage can mean the totals in the two views do not match exactly.
Differences can result from protocol encapsulation, encryption exchanges, retransmissions, and where the client and provider measure traffic. Whether you use Shadowsocks, VMess, VLESS, Trojan, Hysteria2, or WireGuard, do not apply a single fixed “overhead percentage” to every connection. Compare under the same protocol, time window, and network conditions.
3. Set a baseline before resetting: keep statistics easy to review
You do not always need to reset the counters when troubleshooting. The most reliable approach is to record the values before and after a test, then subtract the starting values from the ending values. This preserves accumulated history and makes it less likely that you will mistake an earlier test or background activity for traffic from the current visit.
- Record the conditions. Note the start time, whether you are on Wi-Fi or cellular, the server selected on Home, and the current Global Routing mode.
- Save the baseline. Record Upload and Download for the target server and app. If you want to keep the history, save these readings yourself before resetting.
- Reset if needed. Use the reset control available on the current Data screen. Before confirming, check which counters it affects: the current entry, current view, or all statistics. Do not delete a server as a way to reset counters.
- Confirm the starting point. After resetting, check both views to see which counters changed. If only one changed, use only that view as the starting point for the new period.
- Run one test. Open only the target app, make one repeatable request, then return to Data right away and record the ending values.
Resetting clears local statistics; it does not reset the system’s cellular usage or restore your provider’s plan balance. Quitting the app, disconnecting and reconnecting, or switching servers are not reliable ways to reset counters. If the current screen has no clear reset control, compare before-and-after values instead. There is no need to delete a configuration or reinstall the app just to start from zero.
4. Verify rules with Data: check the destination, match, and outlet
Data helps answer “which category increased?” Connection logs help answer “why did this request go this way?” An increase in a proxy server’s Download figure does not prove that the domain you just opened matched the expected rule; background sync could also have generated proxy traffic. To verify a rule, associate the traffic change with the relevant request.
First, check Home → Global Routing. To verify rules in a configuration, select Config. Proxy and Direct are overall routing modes; test results in those modes cannot prove that a particular rule in the configuration is working. If you use Scene, confirm its actual routing behavior before testing.
The following is a syntax example for the [Rule] section of a configuration, not a complete configuration file. The domain is for demonstration only; for a real test, use a destination you are authorized to access and can reliably test. PROXY must also have a valid outbound target in the current configuration.
[Rule]
DOMAIN-SUFFIX,example.com,PROXY
DOMAIN-SUFFIX,example.org,DIRECT
FINAL,PROXY
- In Config, confirm which configuration is in use and save a copy of its contents before making changes. Rules are checked in order, so first rule out the possibility that an earlier rule already matches the destination.
- On Home, confirm that Global Routing is set to Config and that you are connected to the expected server. Record that server’s starting values in Data.
- Make a new network request in a browser. A cached response may produce little visible network traffic. If needed, use a refresh method supported by the target service instead of simply reopening locally cached content.
- In the connection log, check the destination domain or IP, the matched rule, and the final policy. Port 443 is common for HTTPS, but the port alone does not prove that a connection used a proxy.
- Return to Data and check the corresponding change. When testing DIRECT, do not expect a proxy server’s counter to increase. Whether direct traffic appears as a separate statistic depends on the current view.
If the configuration also contains GEOIP or IP-CIDR, check their position and compare them with the actual destination address. A webpage may contact multiple domains and connect to different IPs; checking only the main domain in the address bar can miss image, API, or video requests. FINAL is the fallback when none of the preceding rules match, not a rule that must run on every request.
In short: use connection logs to verify rules and Data to check traffic changes
If the target connection shows DIRECT while a proxy server’s counter is still increasing, first check for other connections during the same time window. Traffic statistics are useful for troubleshooting only when the destination, timing, and policy are aligned.
5. Narrow down unusual readings based on what you observe
First check whether the counter changes after one clearly defined request, then assess whether the amount seems reasonable. Do not change DNS, servers, protocols, and rules all at once. Change one condition at a time so you can explain the difference. Troubleshoot these four common cases in order.
A webpage loads, but Data barely changes. Why?
First check that the page is not loaded from local cache, then review the Home connection status and Global Routing. If you are watching only a proxy server, DIRECT requests will not noticeably increase that server’s counter. Check the target connection log and make one new, small network request.
Why are the numbers increasing when I am not using my phone?
Email, photo sync, messaging connections, and system tasks can all communicate in the background. Pause sync tasks you can control and observe for one minute to establish an idle baseline; then open one target app and compare the change. You can also check Settings → On Demand for automatic connection behavior, so background traffic after an automatic reconnect is not mistaken for test traffic.
The browser is missing from the app view. Is traffic unaccounted for?
First confirm that the current platform supports the relevant app grouping, then compare the server statistics with the connection log. App identification depends on information provided by the system, so some connections may not be attributed to the expected app name. Do not reinstall the app or change rules just because one app entry is missing.
My provider reports more usage than Data. Which figure should I trust?
Use your provider’s billing figures for your plan balance. First align the billing periods, check whether other devices are using the account, and ask your provider about its billing multiplier, how it charges uploads and downloads, and any reporting delay. Data is for troubleshooting on this device; it does not replace your provider’s bill.
If Upload keeps increasing while Download remains low, first check whether files are being sent, backups are running, or connections are repeatedly retrying. A small amount of upload with almost no download could also mean requests were sent but no valid response arrived. Check connection errors and whether the destination is reachable before concluding that the traffic statistics are faulty.
6. Keep a concise troubleshooting record
When reporting an issue, describing the test conditions and traffic changes is more useful than saying only “the traffic looks wrong.” Get subscription links from your provider; this site does not provide, sell, or recommend proxy servers or subscriptions. Do not include subscription links, passwords, keys, or full server addresses in your troubleshooting record.
- Environment: Device platform, whether you used Wi-Fi or cellular, and the start and end times of the test.
- Routing: The actual Global Routing option, the name of the active configuration, and the order of relevant rules.
- Destination: The redacted destination domain, port, matched rule, and final policy shown in the connection log.
- Readings: Starting and ending Upload and Download values in the server and app views, and whether you reset the counters.
- Actions: The requests made during the test, whether background sync was paused, and whether you changed networks or servers.
Keep the observation window consistent. Verify routing first, interpret the traffic changes next, and only then compare them with system usage or your provider’s bill. This separates routing issues, app attribution limits, and differences in billing methods instead of attributing every discrepancy to the Data page.
Get Shadowrocket and learn how to configure it
Get Shadowrocket for iPhone and iPad from the App Store for a one-time $2.99 purchase. The Mac (Apple silicon and Intel), Windows, Android, and Linux versions are free. Check the App Store listing for iPhone and iPad system requirements. The download page lists ways to get Shadowrocket on each platform, and the setup guide covers importing a subscription and basic routing rules.