firebase-crashlytics
Firebase Crashlytics Skill
This skill defines how to correctly use Firebase Crashlytics in Flutter applications.
When to Use
Use this skill when:
- Implementing crash reporting in a Flutter project.
- Capturing fatal errors, non-fatal exceptions, and async/isolate errors.
- Customizing crash reports with keys, logs, and user identifiers.
- Configuring opt-in data collection or disabling reporting in debug builds.
- Uploading symbol files for obfuscated builds.
1. Setup and Configuration
flutter pub add firebase_crashlytics
flutter pub add firebase_analytics # enables breadcrumb logs for better crash context
Run flutterfire configure to update the Firebase configuration and add the required Crashlytics Gradle plugin for Android.
import 'package:firebase_crashlytics/firebase_crashlytics.dart';
Obfuscated code:
- For apps built with
--split-debug-infoand/or--obfuscate, upload symbol files for readable stack traces. - iOS: Flutter 3.12.0+ and Crashlytics Flutter plugin 3.3.4+ handle symbol upload automatically.
- Android: Use Firebase CLI (v11.9.0+) to upload Flutter debug symbols:
firebase crashlytics:symbols:upload --app=FIREBASE_APP_ID PATH/TO/symbols
2. Error Handling
Configure comprehensive error capture in main() to catch errors from all sources:
Fatal Flutter errors:
void main() async {
WidgetsFlutterBinding.ensureInitialized();
await Firebase.initializeApp();
// Pass all uncaught fatal errors from the framework to Crashlytics
FlutterError.onError = FirebaseCrashlytics.instance.recordFlutterFatalError;
// Catch async errors not handled by the Flutter framework
PlatformDispatcher.instance.onError = (error, stack) {
FirebaseCrashlytics.instance.recordError(error, stack, fatal: true);
return true;
};
runApp(MyApp());
}
Non-fatal Flutter errors: use recordFlutterError instead of recordFlutterFatalError.
Isolate errors:
Isolate.current.addErrorListener(RawReceivePort((pair) async {
final List<dynamic> errorAndStacktrace = pair;
await FirebaseCrashlytics.instance.recordError(
errorAndStacktrace.first,
errorAndStacktrace.last,
fatal: true,
);
}).sendPort);
Caught exceptions (non-fatal):
await FirebaseCrashlytics.instance.recordError(
error,
stackTrace,
reason: 'a non-fatal error',
information: ['further diagnostic information about the error', 'version 2.0'],
);
- Crashlytics only stores the most recent 8 non-fatal exceptions per session — older ones are discarded.
3. Crash Report Customization
Custom keys (max 64 key/value pairs, up to 1 kB each):
FirebaseCrashlytics.instance.setCustomKey('str_key', 'hello');
FirebaseCrashlytics.instance.setCustomKey('bool_key', true);
FirebaseCrashlytics.instance.setCustomKey('int_key', 1);
Custom log messages (limit: 64 kB per session):
FirebaseCrashlytics.instance.log("User tapped on payment button");
User identifier:
FirebaseCrashlytics.instance.setUserIdentifier("user-123");
// Clear by setting to blank string
FirebaseCrashlytics.instance.setUserIdentifier("");
- Avoid putting unique values (user IDs, timestamps) directly in exception messages — use custom keys instead.
4. Performance and Optimization
- Crashlytics processes exceptions on a dedicated background thread to minimize performance impact.
- Fatal reports are sent in real-time without requiring an app restart.
- Non-fatal reports are written to disk and sent with the next fatal report or on app restart.
- Use breadcrumb logs (requires Firebase Analytics) to understand user actions leading up to a crash.
Disable Crashlytics in debug builds:
if (kReleaseMode) {
await FirebaseCrashlytics.instance.setCrashlyticsCollectionEnabled(true);
} else {
await FirebaseCrashlytics.instance.setCrashlyticsCollectionEnabled(false);
}
5. Testing and Debugging
Force a test crash to verify the setup:
FirebaseCrashlytics.instance.crash();
Verification workflow:
- Build and run the app in release mode.
- Trigger the test crash.
- Reopen the app so the crash report is uploaded.
- Check the Firebase Console Crashlytics dashboard within 5 minutes.
- Verify that custom keys, logs, and user identifiers appear on the crash report.
- Verify stack traces are properly symbolicated when using code obfuscation.
6. Opt-in Reporting
By default, Crashlytics automatically collects crash reports for all users.
To give users control over data collection:
- Disable automatic reporting and enable it only via
setCrashlyticsCollectionEnabled(true)when users opt in. - The override value persists across all subsequent app launches.
- To opt a user out, pass
false— this applies from the next app launch. - When disabled, crash info is stored locally; if later enabled, locally stored crashes are sent to Crashlytics.
References
More from evanca/flutter-ai-rules
riverpod
Uses Riverpod for state management in Flutter/Dart. Use when setting up providers, combining requests, managing state disposal, passing arguments, performing side effects, testing providers, or applying Riverpod best practices.
28bloc
Implement Flutter state management using the bloc and flutter_bloc libraries. Use when creating a new Cubit or Bloc, modeling state with sealed classes or status enums, wiring BlocBuilder/BlocListener/BlocProvider in widgets, writing bloc unit tests, refactoring state management, or deciding between Cubit and Bloc.
21effective-dart
Apply Effective Dart guidelines to write idiomatic, high-quality Dart and Flutter code. Use when writing new Dart code, reviewing pull requests for style compliance, refactoring naming conventions, adding doc comments, structuring imports, enforcing type annotations, or running code review checks against Effective Dart standards.
20flutter-app-architecture
Implement layered Flutter app architecture with MVVM, repositories, services, and dependency injection. Use when scaffolding a new Flutter project, refactoring an existing app into layers, creating view models and repositories, configuring dependency injection, implementing unidirectional data flow, or adding a domain layer for complex business logic.
18testing
Write, review, and improve Flutter and Dart tests including unit tests, widget tests, and golden tests. Use when writing new tests, reviewing test quality, fixing flaky tests, adding test coverage, structuring test files, or choosing between unit and widget tests.
16architecture-feature-first
Structure Flutter apps using layered architecture (UI / Logic / Data) with feature-first file organization. Use when creating new features, designing the project folder structure, adding repositories, services, view models (or cubits/providers/notifiers), wiring dependency injection, or deciding which layer owns a piece of logic. State management agnostic.
16