· Advanced · About 7 min read

How to Read Shadowrocket Data Usage: Server and App Statistics

Shadowrocket’s Data screen shows upload and download usage by server or app. Learn what each column counts, when totals reset, and how to tell whether traffic uses PROXY or DIRECT.

At a glance

When checking usage or verifying a rule, note the starting values for a server and an app in Data, then perform an action that you know will use the network. The app row shows which app’s traffic is counted; the server row shows which proxy server’s counter increased. To check whether a connection uses PROXY or DIRECT, also review Global Routing in Home and the active rules in Config.

Uploads and Downloads in Data: Compare Changes, Not Totals

Upload and download are measured from the device’s perspective: data sent by the device counts as upload, and data received counts as download. Opening a webpage generates uploads as requests are sent and connections are established, while the page, images, and other resources generate downloads. It’s normal for both columns to increase during a browsing session; an increase in upload doesn’t mean the app is uploading a file.

2 dimensions
Grouped by server and by app
2 directions
Uploads and downloads tracked separately
2 readings
For the same row and column, subtract the starting value from the ending value

In Data, choose the view you want to check, then note the row name, value, unit, and time. When the test is finished, return to the same row and calculate the change in upload and download separately. If the unit changes, convert the values before subtracting. If the display rounds to a coarse precision, a short request may not change the value shown. Cumulative bytes are not a real-time speed measurement, so a single reading can’t tell you how fast a connection is.

  • Example: The download value for the same server row increases from 12.40 MB to 13.15 MB, an increase of about 0.75 MB on screen. This is a calculation example, not a file size or a billing figure.
  • What’s counted: Use Data to check the traffic recorded by the client. System usage may include other apps or connections that didn’t use the tunnel. Providers may also use their own billing periods and measurement methods.

By Server and by App: Check the Row Name First

A single action in an app can trigger multiple connections. For example, a webpage’s main document, images, and scripts may come from different domains, so an app’s usage can include connections handled by different policies. The server and app views group the same traffic in two different ways; don’t add their totals together.

App requestTraffic entersPolicy checkOutbound connectionUsage is counted

What to Check in Each View

By server

Row name
Note the proxy server you want to monitor before the test
Direction
Record the starting upload and download values for that row separately
Verify
Check the same row after the test and note whether other server rows increased too

An increase in a server row means that server’s counter went up during this period; it doesn’t identify which app caused it.

By app

Row name
Find the app that actually made the test request
Direction
After opening a page, check download first; after sending a file, check upload first
Verify
Compare the app row’s change with the server row’s change over the same period

Use app rows to identify the source of traffic, not as a substitute for per-connection policy records.

If the expected app row isn’t there, first confirm that your action actually made a network request. For example, tapping a link in one app may open the page in a browser, so check the browser’s row too. Reading a document that’s already downloaded or opening a local file in a browser isn’t a useful test.

When to Reset: Record a Baseline Before Testing and Save It Before Clearing

You don’t need to reset usage to check a single request. Keep the existing totals and calculate the difference to run a check without losing prior usage. Clear local statistics only when you need to start tracking from a specific point—for example, to monitor usage for a day after switching your usual server.

  1. Save the previous readings. In Data, note the server and app rows you want to keep, along with the date, time, upload, download, and units. If you’re comparing against your provider’s records, also note the start and end of its billing period.
  2. Check what will be cleared. If the current Data screen offers a clear-statistics option, check whether the confirmation applies to one entry or all entries. Let any downloads in progress finish first so the counters don’t start increasing immediately after the reset.
  3. Set a new baseline. After clearing, check the rows you want to track and confirm that the values have reset or the old entries are gone. If there’s no clear option, record the current totals as your starting values.

Use Data to Check PROXY or DIRECT

First check the mode currently selected in Home → Global Routing, then interpret the changes in Data. Proxy, Direct, Config, and Scene are mode names in the interface; PROXY and DIRECT are policy values used in rules. With Scene, confirm which settings are active under the current network conditions. With Config, check the active configuration, not just a saved file.

  1. Limit the test scope. Pause other downloads or sync tasks you can control. Use one app to download an HTTPS file from a host you can identify. Before starting, note the upload and download values for the app and server rows, then confirm the new download finishes.
  2. Check rule order. In Config mode, check whether an earlier rule matches the target host. For example, DOMAIN-SUFFIX,example.com,DIRECT matches example.com and its subdomains; FINAL,PROXY is used only if no earlier rule matches. If the download is redirected, check the domain that ultimately serves the file too.
  3. Compare changes and cross-check. You can attribute a proxy connection to the test only when the app row and expected server row both increase during the same test period and the target connection’s policy record points to the proxy. If the download completes, the expected server row doesn’t increase, and the connection record shows DIRECT, that connection is confirmed as direct. Without a connection record, treat counter changes as clues only.

A single webpage can use both PROXY and DIRECT connections. Check the policy for each target connection rather than classifying the entire page or app as one or the other. If a policy in the configuration selects a server automatically, confirm which server the connection actually used; the selected item shown in Home alone isn’t enough.

Unexpected Data Readings: Common Questions

Troubleshoot by symptom

The webpage loaded. Why didn’t the server’s download counter change?

First rule out cached content: try a test file that hasn’t been downloaded before, wait for the download to finish, then check Data. If the value still hasn’t changed, check Global Routing and the target connection’s policy, and make sure you’re looking at the row for the server used in this test. An unchanged server row alone doesn’t prove the connection used DIRECT.

I haven’t opened a webpage. Why is a server row still increasing?

Pause background sync and downloads you can control, take two readings, and see whether the increase stops. Then compare the app rows over the same period. Background requests and connectivity checks can also increase usage; a change in a server row alone doesn’t identify a foreground app as the source.

I switched servers. Why does the old server still show usage?

Previous totals may remain visible. Record the values for the old and new server rows, then make another test request. If the old row doesn’t increase, the displayed value is from earlier usage. If it keeps increasing, check for existing connections still transferring data or other rules that still select that server.

The provider’s usage doesn’t match Data. Which should I check?

Use Data to investigate traffic changes on this device, and your provider’s records to check your remaining plan allowance. Before comparing, align the time period, units, and devices included, then check whether the provider combines usage from other devices on the same account.

Download Shadowrocket