Customer Cases
Pricing

How to Find and Fix Memory Leaks in Android Apps (Ultimate Guide)

When the tester tells you that your XXActivity has leaked, how do you confirm whether it has really leaked? After confirming the leak, how do you locate the problem causing the memory leak?

Introduction

When the tester tells you that your XXActivity has leaked, how do you confirm whether it has really leaked?

After confirming the leak, how do you locate the problem causing the memory leak?

In daily Android development, the hardest hit area of memory leaks is Activity. It is believed that these two issues have been encountered by every Android developer. When encountering such problems, we generally use our killer trick: Dump Java Heap and then perform static analysis on the GC chain with MAT. However, today I want to take a different approach and solve this problem from a simpler perspective.

Confirm the leak

First, let's take a look at an abstract Activity pseudocode:

public class LeakActivity extends Activity {

    ComplexLogicA; // Complex business logic code

    @Override
    protected void onCreate(Bundle savedInstanceState) {
        super.onCreate(savedInstanceState);

        ComplexLogicB; // Complex business logic code
        LeakLogic; // Business logic code causing the leak
        ComplexLogicC;
        ComplexLogicD;
    }

    @Override
    protected void onResume() {
        super.onResume();

        ComplexLogicE;
        ComplexLogicF;
        ComplexLogicG;
    }

    OtherComplexLogin... // Other business logic code
}

 

 

If we want to confirm whether this Activity has a leak, we just need to override the finalize method of the Object and add a Logcat print statement inside:

public class LeakActivity extends Activity {

    ComplexLogicA; // Complex business logic code

    @Override
    protected void onCreate(Bundle savedInstanceState) {
        super.onCreate(savedInstanceState);

        ComplexLogicB; // Complex business logic code
        LeakLogic; // Business logic code causing the leak
        ComplexLogicC;
        ComplexLogicD;
    }

    @Override
    protected void onResume() {
        super.onResume();

        ComplexLogicE;
        ComplexLogicF;
        ComplexLogicG;
    }

    @Override
    protected void finalize() throws Throwable {
        super.finalize();

        Log.d("", "====LeakActivity has been recycled!");
    }

    OtherComplexLogin... // Other business logic code
}

 

Then run your project, open this Activity, press the back button to exit the Activity, and then force a GC operation through the IDE:

(Android Studio)

(Eclipse)

Next, check the Logcat to see if there is a corresponding print statement. This will confirm whether the Activity has a memory leak: if there is a print statement, there is no memory leak; if there is no print statement, there is definitely a memory leak!

Locate the cause of the leak

Locating the cause of the leak is a relatively simple and crude process of elimination. First, comment out all the complex business logic until the memory leak no longer exists:

public class LeakActivity extends Activity {

//    ComplexLogicA; // Complex business logic code

    @Override
    protected void onCreate(Bundle savedInstanceState) {
        super.onCreate(savedInstanceState);

//        ComplexLogicB; // Complex business logic code
//        LeakLogic; // Business logic code causing the leak
//        ComplexLogicC;
//        ComplexLogicD;
    }

    @Override
    protected void onResume() {
        super.onResume();

//        ComplexLogicE;
//        ComplexLogicF;
//        ComplexLogicG;
    }

    @Override
    protected void finalize() throws Throwable {
        super.finalize();

        Log.d("", "====LeakActivity has been recycled!");
    }

//    OtherComplexLogin... // Other business logic code
}

 

Next, enter and exit the Activity again, trigger GC, and confirm that the Activity leak no longer exists. Then, add the business logic back one by one until the leak reappears.

In this way, we can 100% locate the cause of the leak.

Key point

The knowledge point used here is the finalize method in the Object class in Java. When GC is ready to recycle a Java Object (all Java objects are subclasses of Object), GC will call the finalize method of this Object. This method is similar to the destructor in C++, and its original intention is to allow you to recycle some resources that are no longer needed (mainly for native resources). In fact, in daily Java development, it is not encouraged to rely on this method to implement recycling logic, because if you heavily rely on finalize, it may cause memory leaks. However, in our case, it is just used as a basis for whether it has been recycled, which is acceptable.

Conclusion

Although this method may seem simple and crude, it is indeed a convenient self-test method for developers. Without using other tools, you can quickly determine whether the newly added Activity in your current requirements has a leak. Consider it as a way to reduce the number of bugs in your code!

Latest Posts
1WeTest at GDC 2026: AI Automated Testing Ushers in a New Era of Game Quality WeTest at GDC 2026 showcases a revolutionary AI Automated Testing Solution that transforms game quality assurance. Learn how WeTest's AI Test Agent Platform enables scalable quality production through computing power, delivering controllable, reproducible, and intelligent testing capabilities.
2Precision Testing in Practice: A Fund Team's Journey from Experience-Based to Data-Driven Quality Assurance Learn how Shenwanhongyuan Securities implemented precision testing to reduce regression testing by 67%. This technical guide covers JaCoCo implementation, method-level code mapping, and intelligent test case recommendation for financial services applications.
3How to Do Performance Test Monitoring: Key Metrics & Tuning Tips Learn how to do performance test monitoring effectively. Discover key metrics (RT, TPS, IOPS), identify CPU/memory/database bottlenecks, and follow step-by-step tuning tips for stable, efficient systems.
4The Ultimate Guide to AI Agent Performance Testing Learn comprehensive AI Agent performance testing strategies, environment setup, tool selection, and optimization techniques. Master how to ensure stability and efficiency in production.
5LLM Security Testing in ToB Scenarios: A Practical Guide & Framework Explore the unique security risks of LLMs in ToB scenarios, including prompt injection and system prompt leakage. Learn about the 'llm-safe-test' framework and how to automate safety judgment for enterprise AI applications.