This guide uses the Data page on iPhone and iPad as an example. It’s for anyone checking whether a download went through the selected server or tracking what caused traffic to increase. First, distinguish cumulative data from live speed, then choose the right stats view. When troubleshooting, compare before-and-after values on the same row—you don’t need to reset the counters first.
Uploads and downloads: Check the units, then compare the change
The directions are relative to your device: uploads are data sent by your device, and downloads are data it receives. Even loading a web page involves uploads, such as sending a request; uploading a file can also involve downloads, such as receiving the server’s response. So it’s normal for both counters to increase during one activity. The Data page shows usage Shadowrocket records within the selected stats view. It doesn’t show all Wi-Fi or cellular usage on your device and isn’t intended to match your carrier bill.
Cumulative values shown in KB or MB answer “how much data has been transferred?” A speed value with /s answers “how much data is transferred per second?” For example, if the download value on a row changes from 12.4 MB to 15.6 MB, the download increase over that period is about 3.2 MB. Convert both readings to the same unit before comparing them. Values are rounded, so a short request may transfer data without changing the last displayed digit.
To check one activity, note the starting value
Before and after the activity, check the same view, row, and direction. Don’t attribute an existing cumulative value to a page you just opened, and don’t estimate speed by dividing the total by the time spent on this activity.
By server or by app: choose the right stats view
The two views group traffic differently. To check whether traffic is recorded under the expected proxy server, start with the server view. To see which app entries increased over the same period, use the app view. The suggestions below assume you’re checking a proxy route.
Traffic by Server
RecommendedNote the selected server’s name on Home, then return to Data and check the change for its entry. After switching servers, the previous entry may retain its cumulative value; that doesn’t mean the old server is still handling new connections. If a rule specifies a different policy, the server selected on Home alone won’t tell you where every request is routed.
Use this view to check whether a proxy server has new traffic and compare readings before and after switching servers.
Traffic by App
Compare each app entry’s upload and download changes over the same period, rather than looking only at cumulative rankings. The app view shows traffic Shadowrocket records and categorizes in that view. An app missing from the list isn’t necessarily offline.
Use this view to find what’s driving increased usage and distinguish foreground activity from background activity in other apps.
Verify Data readings with a single download
Choose a test file with a known size that hasn’t been cached, such as a file around 5 MB. Pause any sync or automatic downloads you can control during the test, and stay on the same network. If you reuse a file, make sure the browser starts a fresh download rather than opening a local copy.
Check the connection and routing
Open Home, confirm that the connection is established, and note the selected server and current Global Routing option. If you use Config, check that the active configuration routes the test request through the expected policy. A request matched to DIRECT won’t verify the selected proxy server’s counter.
Record the readings before downloading
Open Data, find the target entry in the server view, and note its upload, download, and units. If you also want to check app categorization, note the corresponding entry in the app view. If the entry isn’t there yet, record it as “not shown”—don’t treat it as a measured zero.
Run one download
Switch to the test app, start one download, and wait for it to finish. Don’t switch servers, open other test pages, or run Connectivity Test after recording the baseline; those requests could affect the readings for the same period.
Return and compare the readings
Go back to Data and check the same row in the same view. Subtract the baseline from the reading after the download. If the target server’s download value increased, check how the traffic is categorized in the app view. File size is a useful reference, not a value the counter must match exactly: response headers, extra requests, and background transfers can all affect the result.
Resetting counters: Check the scope, then start a new baseline
Resetting counters starts a new observation period for the client’s usage data. It doesn’t reset cellular usage or reverse traffic already counted by your carrier. If you only need to compare usage before and after an activity, keep the existing values and calculate the difference. Reset counters from Data only when you need to start tracking from scratch.
- Before clearing: Note the time, current view, and any cumulative values you need to keep. If you want a stable baseline, finish any transfers first so the readings don’t start increasing as soon as you confirm.
- When confirming: Follow the confirmation prompt on screen and check whether the action affects the current entry or a wider range. Don’t infer the scope from the view you’re in. If the prompt doesn’t specify the scope and you need to keep other records, cancel and compare the readings instead.
- After clearing: Return to the same view and check whether the entry has reset to zero or disappeared. To verify the new starting point, run another controlled download. If the connection is still in use, new nonzero readings don’t mean the reset failed.
Unexpected readings: Common causes and checks
A webpage loads, but the server counter doesn’t increase. Why?
First check the connection status and Global Routing on Home. If the route is Direct, or the test domain matches DIRECT in Config, it’s normal for the selected proxy server’s counter not to increase. If the request should use a proxy, check whether the configuration routes it through a different server or policy, then test again with a file that hasn’t been cached.
I opened just one app. Why are other rows increasing?
Stop testing and note the Data readings, then check the same rows again after about 30 seconds. If they increase while idle, pause any sync, downloads, or media playback you can control and repeat the check for a similar period. An app being in the foreground doesn’t mean all traffic during that period came from it.
Why don’t Data readings match the system’s cellular usage?
First align the start times, then confirm that the test used cellular data throughout. System cellular statistics may include traffic that didn’t pass through Shadowrocket, while Data may also record traffic over Wi-Fi. The two screens cover different traffic, so note each one’s before-and-after values separately on the same network instead of comparing historical totals.
I switched servers, but the old server’s counter is still there. Is that normal?
A retained cumulative value doesn’t mean traffic is still flowing through that server. Note the starting values for both the old and new entries, finish the current download, then start a new one. If the old entry keeps increasing, check whether the new request reused an existing connection and whether the active policy in Config or Scene still points to the old server.