Crashes on Android

Android App Crashing: Halt the Rollout, Read Google Play's Numbers, Then Fix the Cause

The one-star reviews arrive in clumps, the way bad news does. "Opens and closes." "Useless since the update." Reviews tell you something is wrong. Google Play's own numbers tell you how wrong, on which phones, and how long you have before it starts costing you visibility.

November 20, 2026 · 10 min read

acmefieldnotes app · crash rate by day
Version 4.2 on a staged rollout
Rollout halted, then the fix went to a slice first
User-perceived crash rate, dailyHalted
Day 14.2 reaches 20% of users; rollout haltedDay 12
Illustrative: sample crash trend for a fictional field service company's Android app

The short answer

When an Android app starts crashing, open Android vitals in Google Play Console and find the version, device and stack trace behind the spike. If the bad release is in a staged rollout, halt it so no more users receive it. Then fix the cause and ship a new release. As of November 2026, Google's bad behavior threshold for user-perceived crash rate is 1.09% overall and 8% per phone model.

On this page
  1. Google Play keeps score whether you look or not
  2. Sixty minutes after the bad reviews begin
  3. A staged rollout is a fire door, and it only closes one way
  4. What the stack trace is trying to tell you
  5. Crashes that arrive without a new release
  6. The hotfix that makes Tuesday worse than Monday
  7. Fix it in-house, or bring us in

Google Play keeps score whether you look or not

Google measures your app's stability on real phones and acts on what it sees. Its Android vitals documentation, read in November 2026, names the core vitals as "user-perceived crash rate", "user-perceived ANR rate", "excessive partial wake locks", "memory usage" and "bitmap memory usage", and states that "Your app's core vitals affect your app's visibility on Google Play." Google's ANR guide explains the acronym: when an app's UI thread "is blocked for too long, an 'Application Not Responding' (ANR) error is triggered," and a foreground app shows a dialog that lets the user force quit it.

Google Play bad behavior thresholds for the two stability vitals, as recorded in Android vitals documentation, as of November 2026
VitalOverall (average across devices)Per phone modelPer watch model
User-perceived crash rate1.09%8%4%
User-perceived ANR rate0.47%8%5%

The consequence is spelled out: "If your app or game exceeds a bad behavior threshold, Play may reduce the visibility of your title. Play may also show users a warning on your store listing." Google says Play uses "the last 28 days of data to assess your app's quality," and that warnings go away when the 28-day average improves, as of November 2026. A crash spike today shows up in that window for weeks, which is why speed in the first hour matters more than elegance.

Google also flags emerging issues, problems "affecting devices for over 7 days for crashes and ANRs. This gives you 21 days to address them," as of November 2026. Memory, bitmap memory and code optimization thresholds come later; Google says apps over those "may see store visibility impact starting from February 2027."

Sixty minutes after the bad reviews begin

Work from the numbers, not the reviews. Reviews are loud, late and vague. Play Console is quiet, early and specific. Do these in order and write down what you find at each step.

  1. 1

    Open Android vitals

    In Play Console, find the crash and ANR rates and look for the day the line jumped. Note which app version, which Android versions and which phone models carry the crashes.

  2. 2

    Halt the staged rollout if there is one

    If the bad version is still rolling out, halting it stops new users from getting it. This is the single most valuable click of the hour; the next section explains what it does and does not do.

  3. 3

    Read the top stack trace

    Google's developer docs say "You can view crash stack traces in Android vitals." Copy the top cluster's trace into a document, with the exception type and the first line of your own code in it.

  4. 4

    List what changed, inside and outside the app

    The release itself, but also server deploys, API changes, a rotated or revoked key, a new third-party SDK version, a remote config flag. Apps crash when the world around them changes too.

  5. 5

    Try to reproduce it

    On a phone matching the worst-hit model and Android version, follow the steps users describe. A crash you can make happen on purpose is half fixed.

  6. 6

    Tell users once, plainly

    Reply to the reviews with one honest line and a fix-in-progress note. Do not promise a release date you have not earned yet.

If the crash also hits your iPhone app, the cause is probably shared: a server, an API or a login service. Our iOS app crashing page covers Apple's side, and the app login broken page covers the case where the crash sits right behind the sign-in screen.

A staged rollout is a fire door, and it only closes one way

Google's staged rollout help page, read in November 2026, describes the feature simply: "With a staged rollout, your update reaches only a percentage of your users, which you can increase over time." When you find a problem, you can halt the rollout "to help minimize the number of users who experience the issue." Google adds that the percentage "won't increase automatically," so a rollout only grows when someone grows it.

Bright blue duotone of a smartphone with a blank screen lying at an angle, surrounded by solid circles, triangles and bars in blue and white
A halted rollout protects the users who have not updated yet; it cannot reach back to the ones who have.

If you shipped the update to everyone at once, there is no door to close, and the fix becomes a race. It is a lesson worth learning only once: the next release goes out to a slice of users first. If the trouble started with a release of your web or server code rather than the app, the failed software update page covers rolling that back.

What the stack trace is trying to tell you

A stack trace looks like noise to an owner and like a confession to a developer. Android's guide to crashes, read in November 2026, says an app "crashes whenever there's an unexpected exit caused by an unhandled exception or signal," and that a trace shows two critical things: the type of exception thrown and the section of code where it was thrown. The exception type, Google says, "is usually a very strong hint as to what went wrong."

NullPointerException

Code reached for something that was not there. Google's guide says "Null pointer exceptions are the largest cause of app crashes on Google Play." A server that suddenly sends an empty field is a classic trigger.

OutOfMemoryError

The phone ran out of memory for the app. Google suggests testing on an emulator set up with low memory, because development machines have more to spare than users' phones.

UnknownHostException

A network call failed. Google's guide says network exceptions usually should not be treated as errors but as normal conditions, and suggests airplane mode to reproduce them.

A crash with no app on screen

Google notes that "An app doesn't need to be running in the foreground for it to crash." Background components can crash it too, which confuses users who were not even using it.

Traces with no readable names

If native traces lack class and function names, Google says you may need to generate a native debug symbols file and upload it to Play Console. Unreadable traces slow every fix after this one.

You do not have to understand the code to get value here. Copy the exception name and the first line that mentions your app's package, and you have given any developer, including us, the starting point that saves the first hour of their time.

Crashes that arrive without a new release

The most confusing Android emergencies happen when nobody shipped anything. The app on users' phones is the same file it was last week. Something it depends on changed instead. Our reasoning, not a Google statement: an app that trusts its server can crash the moment the server's answer changes shape.

  • A backend deploy renamed or removed a field the app expects, and the app meets a null it never planned for.
  • An API key was rotated or revoked, a certificate expired or a third-party account lapsed, and the app does not handle the failed call gracefully. Our API integration repair page covers the 401, 403 and 429 responses behind many of these.
  • A remote config flag or feature switch was flipped for everyone instead of a test group.
  • A third-party SDK, such as analytics or ads, misbehaves on a new Android version or a specific phone model.
  • The login provider changed or went down, and the app was never written to show an error instead of closing.

Check these before blaming the last release. The good news is that a server-side cause can often be fixed without a new app version at all, which means no store review in the path. The web app down checklist covers the server side.

The hotfix that makes Tuesday worse than Monday

Nothing spreads a crash faster than a fix pushed to everyone at once.
  • Shipping an untested hotfix to 100% of users. If it is wrong, you have two bad versions in the field. Stage the fix too.
  • Guessing at the cause from reviews instead of reading the stack trace. Reviews describe symptoms on one phone; vitals describe the pattern.
  • Ignoring ANRs because they are not crashes. Google counts the user-perceived ANR rate as a core vital with its own threshold.
  • Removing crash reporting to make the dashboard quieter. The crashes keep happening; you just stop seeing them.
  • Rebuilding with a different signing setup or account in a panic. If the person with Play Console access is gone, fix the access problem properly first.

Lost access is its own emergency. If the developer who owns the Play Console account has vanished, read the developer disappeared page before anything else; nobody can halt a rollout from outside the account. If Google rejects the fixed update, its policy page says plainly that "A new app or app update submitted for review will not be made available on Google Play" until the issue is resolved, as of November 2026.

Fix it in-house, or bring us in

Fix it in-house when your developer is reachable, the stack trace points to one obvious line, and the rollout is halted. That is a normal day in app development, and you do not need an emergency firm for it. Call us when nobody can read the trace, when crashes appear without a release, when the original developer is gone, or when the fix keeps failing review. We respond the same business day (hours on our contact page).

Android crash repairs are from $575, depending on complexity, with a written estimate before work starts; emergency pricing explains what moves that number. We will need access to Google Play Console and the source code. We promise no store approval and no fix time, because nobody honestly can. The full mobile offer is on our emergency app repair page, the wider service is fixing broken mobile apps, and apps that need rebuilding fall under mobile app development. See every scenario on the emergency repair hub, our published work, and who we are: a San Antonio studio building for the web since 1999, at (210) 346-0848.

Questions owners ask in an emergency

Google's Android vitals documentation sets the bad behavior threshold for user-perceived crash rate at 1.09% overall, 8% per phone model and 4% per watch model, as of November 2026. For user-perceived ANR rate it is 0.47% overall, 8% per phone model and 5% per watch model. Exceeding them can reduce visibility and add a listing warning.

Google's staged rollout help page describes halting, not rolling back. Halting stops new users from getting the version, but users who already received it remain on it. Google's advice for a bad bundle is to create and roll out a new release with a fixed app bundle. Plan to fix forward, and stage the fix.

Google says Play uses the last 28 days of data to assess quality, and warnings disappear when the 28-day average improves, as of November 2026. For emerging crash and ANR issues affecting devices for over 7 days, Google says you have 21 days to address them.

Google's developer guide says you can view crash stack traces in Android vitals in Play Console. Each trace shows the exception type and where in the code it was thrown. If native traces lack readable names, Google says you may need to upload a native debug symbols file to Play Console.

Something the app depends on probably changed: a server response, an API key, a remote config flag, a third-party SDK or a login provider. Check your backend deploys and integrations first. A server-side cause can often be fixed without a new app version, so no store review is needed.

DataCram's mobile app repairs are from $575, depending on complexity, with a written estimate before work starts. We respond the same business day and need access to Google Play Console and the code. We do not promise fix times or Google approval.

Keep reading

Free · No obligation

Get the crashing app looked at today

Tell us the app, the version that started crashing and what Android vitals shows. We reply the same business day with what we would check first and a written estimate.

Which of these sounds like you? (pick any)

Rather talk now? Call (210) 346-0848.

CallGet a quote