The screen going dark is not itself the fault
You start recording a run on your phone, put it away and find a long straight line on the map at the finish where the actual turns should be. Or the recording simply ends in the second minute after starting. If this follows locking the display, suspicion is understandable, but the screen going dark is not an explanation in itself. You need to establish the operating state in which the app tried to continue collecting location data.
To the user, a phone in a pocket is “working in the background”, but Android’s technical categories are more precise. According to Google’s guidance, a location feature using an appropriate foreground service may still count as foreground location access when the app window is no longer visible or the display is off. A locked screen therefore does not automatically mean every such running app requires permanent background permission. Source: Android location-access categories.
This article troubleshoots using Android examples. On another system, follow the platform’s and specific app’s guidance rather than looking for the switch names used here. The method transfers, however: first distinguish the type of interruption, then examine the required access and the app’s recording mode. The aim is a predictably functioning setup of your own that does not require the screen to stay lit throughout a run.
Three separate events can produce a similar map
In the first case, recording really stops. Returning to the app, the workout is no longer running, or the saved activity ends at the suspected interruption. In the second, its timer continues but no usable location points accumulate. In the third, a recording exists but later display or processing conceals a gap. Do not describe all these possibilities with one phrase, “GPS did not work”.
Include what you saw when reopening the app in your observations. Did timing continue? Was there a warning? Was the previous workout screen still there? Did total distance also stop, or did only the map draw strangely? Apps do not display these in identical ways, so exact wording and time are more valuable than a diagnosis chosen beforehand. If stopping to check, do so somewhere safe; you need not inspect menu after menu while running beside a road.
Do not automatically equate losing internet access with location recording stopping either. Loading map tiles, uploading an activity and collecting location data may be separate app functions. Your program’s documentation determines their dependencies. A recording synchronising only later does not mean location data were absent meanwhile. Conversely, a working mobile-data icon does not prove the running app recorded throughout.
What does the foreground location service do?
Android documentation describes a location-type foreground service for longer-running location tasks. It has developer requirements, and appropriate permission must also be provided. From Android 14, the document details declarations and permissions for specific service types. The app developer implements these; users cannot replace missing app-side functionality with a hidden system switch. Source: Android foreground-service types.
For users, the important point is to follow the app’s prescribed starting sequence. Starting the workout and then locking the device may be a different operating situation from a new location feature trying to start from another app while already in the background. Do not assume the two cases are identical on every system version. If the developer requests starting the workout in the foreground, test that sequence, then examine continued operation in your pocket.
A notification indicating app operation is a useful clue but does not itself prove complete data recording. Even with a visible service indication, inspect the saved route. If a previously present status indicator disappears at the interruption, note this. Keep observations precise in the test log: “the notification disappeared then”, rather than “the system definitely did this or that”. The latter claim may require app logs or developer investigation.
Source: characteristics of location permission.Source: Android background location.
If changing a permission only for a test, record the previous state and the purpose of the change. After success, examine whether the modified level is genuinely necessary for the desired function. Do not accidentally leave access you did not intend to make permanent. A good result is more than one working recording: it is an understood setting whose functional and privacy consequences you accept. A test permission should not become a forgotten decision.A short test should reveal the effect of locking
Choose a short, familiar loop with several clear turns and safe places to stop. Do not make the next long race your first test. Begin with the app’s documented starting sequence, note the start time, then lock the display at a known point and put the phone in its usual location. Make the next check at a preselected calm spot. This allows comparison of display state and recording time.During the test, do not change carrying position, power mode, permission and app version together. Changing everything at once gives a later success too many possible explanations. Record the existing state first, then examine one justified setting according to that app’s official guide. The environment can change between tests too, limiting what your own result can establish. Even so, identifying a substantial recurring interruption can help considerably.
At the end, save the workout first, then review the complete recording. The beginning may have worked well even if a later gap developed. Observe turn locations, continuity of the time series and when the discrepancy begins. If the app provides an exportable file, retain the original. A screenshot is a useful reminder but may not show location-point timestamps or the actual gap duration.
A verifiable screen-lock test
Documented start
Start with the app’s own sequence and known settings.
Marked locking time
For a numerical example, suppose the two surviving points are 600 metres apart and six minutes elapsed between them. This would yield a calculated connecting pace of 10:00 minutes per kilometre. If you actually ran 1000 metres with a detour, the same time would give 6:00. The incomplete point series alone cannot choose the real one. This example does not describe a known app’s algorithm; it shows why a line drawn over missing data should not be read as actual performance.
Power management: use your own app’s guide as the referenceThe phone’s power mode and settings governing app operation may form part of the investigation, but their names and effects can vary between devices. Do not disable every system protection based on an old forum list for another manufacturer. Find your running app’s current support for the particular phone or system. If documentation recommends a specific exception, test it purposefully and record the outcome.
Google’s location-permission description specifically notes that, under certain power-saving operation, location updates can continue in the foreground or through a foreground service even with the display off. “Power saving always stops the run” is therefore too general. The actual system state and app implementation matter together. Record the exact power-mode name and observed result in your test, rather than extending an explanation to every phone.
Source: location use in power-saving states.
If a setting improves operation, avoid accidentally reversing it during the next system update or app reinstall. A short note of the working setup preserves the decision. One good result is not an eternal guarantee, however: after a significant update, recheck your use situation on the same short loop. A targeted test costs less than discovering missing recording after a full marathon.
Open illustration at full size
A runner reviewing training-log data writes a separate note about an uncertain measurement.Enlarge image
AI-generated editorial illustration of clearly labelling suspect sensor data; not a documentary photograph.
Enlarge imageWhat should you send support?
Give the exact phone model, system and app versions, actual location permission, starting sequence and approximate interruption time in the fault report. Describe whether the issue coincided with locking the display or opening another program. If this is still only a hypothesis, call it that. “The same fault in three consecutive short tests” is more informative than “it never works” when some occasions produced complete recordings.
Keep evidence proportionate to the question. Settings images cleared of personal information and a test in a neutral location are often sufficient first steps. If an original file is requested, supply the necessary portion through an official private channel and consider whether it includes your home location or other sensitive information. Publicly sharing your full movement history is not a prerequisite for meaningfully demonstrating a background-operation problem.
If the developer requests app-side logs or another test, track changes individually. User settings cannot repair every code-side operating fault, so failure is not necessarily your fault. A precise test can also show that a recurring problem remains despite supported settings. You can then pass the documented case to the developer and meanwhile choose another previously tested recording solution for an important event.
A working recording does not require constantly watching the display
With a reliably configured running app, know your own supported workflow: how you start, when you lock the device, where you carry it and how you check the saved recording. There is no need to turn this into a long technical list before every run. A few targeted checks and a short test after an update may be enough to keep your setup predictable. The method serves your routine rather than creating another screen-dependent task.
Label the earlier incomplete workout in the log with the fault type. For example, “timing retained, route missing for six minutes” or “app recording ended early”. You will then avoid comparing it later with a good recording as though it were a complete, precise pace test. The activity happened even if the phone failed to preserve all of it. Honestly marking data quality helps separate the experience of training from the possibility of numerical analysis.
An interruption when the screen is locked is therefore an operating situation to observe rather than the name of one switch. Permissions, the app’s recording mode, system version and saved data provide the answer together. Examining them separately avoids blindly granting ever broader access. The result is a deliberately tested setting that lets you put the phone away and focus on the route, surroundings and your own running on the next loop.
Sources
- Android: location permissions
- Android: foreground-service types
- Android: background location
- GPX 1.1: track segments
How Futószív articles are prepared: our editorial guidelines.


