Debugging scenarioMedium~10 minGC and Lifetime / 7 of 8

Interview question

A component subscribes to an event and memory grows over time. How can the subscription cause a leak?

How can event subscriptions cause memory leaks in .NET?

Shows how publisher-to-subscriber references can retain objects and how to clean up event-based lifetimes.

1Say itAnswer out loud first

2CompareCompare with Core answer

3RefineRefine with Stronger and Advanced

TL;DR

In .NET, an event subscription usually means the publisher holds a reference to the subscriber delegate. If the publisher lives longer than the subscriber, that reference can keep the subscriber alive. For example, a short-lived view model or service that subscribes to a singleton event can remain in memory after it is logically done. I would unsubscribe at the right lifecycle point or avoid the long-lived event path if ownership is unclear.

Full answer below

Practice focus

Reference ownership, event publisher lifetime, unsubscribe patterns, weak events, and leak diagnosis.

Answer first

Answer path

Core answer first, then checkpoints toward a stronger answer.

Start
Core answer
Start here. This is the direct spoken answer to practice first.
Why this question matters
An event publisher holds subscriber delegates. When the publisher outlives a subscriber, that reference can retain the subscriber and its object graph long after the logical operation or UI lifetime has ended.
In .NET, an event subscription usually means the publisher holds a reference to the subscriber delegate. If the publisher lives longer than the subscriber, that reference can keep the subscriber alive. For example, a short-lived view model or service that subscribes to a singleton event can remain in memory after it is logically done. I would unsubscribe at the right lifecycle point or avoid the long-lived event path if ownership is unclear.
Checkpoint
Stronger answer
Optional upgrade with examples, trade-offs, and sharper phrasing.
I would explain it as a reference graph problem. The subscriber may disappear from the UI or request flow, but the publisher still has a delegate pointing to it, so the GC sees it as reachable. The fix depends on the design: unsubscribe in Dispose, use a scoped publisher, use weak-event patterns in UI-style code, or replace broad events with explicit callbacks that have a clear owner. I would also check lambdas because anonymous handlers are easy to subscribe and hard to remove correctly.
Finish
Advanced answer
Optional deeper angle for production judgment and nuance.
In production I would verify the root path before changing patterns across the codebase. If a heap dump shows many old subscriber instances rooted through the same publisher event, I would fix that lifecycle boundary and add a regression test around disposal or unsubscribe behavior. I would not replace every event with weak references, because weak patterns can hide ownership bugs and make behavior harder to reason about. Prefer matching publisher and subscriber lifetimes or using a small ownership wrapper that makes unsubscription automatic and testable.

Review signals

Use these to catch weak phrasing, shallow reasoning, and risky claims.

Common mistakesCandidateCommon mistakesMistakes that weaken the answer or signal shallow understanding.Open details

Details

Subscribing short-lived objects to singleton events.
Using anonymous handlers that cannot be removed easily.
Assuming the subscriber disappears when it leaves the UI or request.
Adding weak references without understanding ownership.

Red flagsCandidateRed flagsSignals that an interviewer may treat as risky.Open details

Details

Cannot describe the publisher-to-subscriber reference.
Does not ask which object lives longer.
Only suggests increasing memory limits.
Ignores lambdas and anonymous event handlers.

Follow-up prompts

Quick interviewer prompts inside this drill.

Follow-up questionsCandidateFollow-up questionsLikely next questions after the main answer.Open details

Details

Treat these as the next interviewer turn. Answer one out loud, then save the gap in your practice notes.

Who holds the reference after an event subscription?
When would Dispose unsubscribe?
Why can anonymous handlers be risky?
How would a heap dump reveal this leak?

Continue the section

Next in GC and Lifetime: Singleton Capture

Step 8 of 8

Why is it risky for a singleton to capture scoped services or request data?

Opening drill

Debugging scenarioMedium~10 minGC and Lifetime / 7 of 8

Interview question

A component subscribes to an event and memory grows over time. How can the subscription cause a leak?

How can event subscriptions cause memory leaks in .NET?

Shows how publisher-to-subscriber references can retain objects and how to clean up event-based lifetimes.

1Say itAnswer out loud first

2CompareCompare with Core answer

3RefineRefine with Stronger and Advanced

TL;DR

Full answer below

Practice focus

Reference ownership, event publisher lifetime, unsubscribe patterns, weak events, and leak diagnosis.

Answer first

Answer path

Core answer first, then checkpoints toward a stronger answer.

Start
Core answer
Start here. This is the direct spoken answer to practice first.
Why this question matters
An event publisher holds subscriber delegates. When the publisher outlives a subscriber, that reference can retain the subscriber and its object graph long after the logical operation or UI lifetime has ended.
In .NET, an event subscription usually means the publisher holds a reference to the subscriber delegate. If the publisher lives longer than the subscriber, that reference can keep the subscriber alive. For example, a short-lived view model or service that subscribes to a singleton event can remain in memory after it is logically done. I would unsubscribe at the right lifecycle point or avoid the long-lived event path if ownership is unclear.
Checkpoint
Stronger answer
Optional upgrade with examples, trade-offs, and sharper phrasing.
I would explain it as a reference graph problem. The subscriber may disappear from the UI or request flow, but the publisher still has a delegate pointing to it, so the GC sees it as reachable. The fix depends on the design: unsubscribe in Dispose, use a scoped publisher, use weak-event patterns in UI-style code, or replace broad events with explicit callbacks that have a clear owner. I would also check lambdas because anonymous handlers are easy to subscribe and hard to remove correctly.
Finish
Advanced answer
Optional deeper angle for production judgment and nuance.
In production I would verify the root path before changing patterns across the codebase. If a heap dump shows many old subscriber instances rooted through the same publisher event, I would fix that lifecycle boundary and add a regression test around disposal or unsubscribe behavior. I would not replace every event with weak references, because weak patterns can hide ownership bugs and make behavior harder to reason about. Prefer matching publisher and subscriber lifetimes or using a small ownership wrapper that makes unsubscription automatic and testable.

Review signals

Use these to catch weak phrasing, shallow reasoning, and risky claims.

Common mistakesCandidateCommon mistakesMistakes that weaken the answer or signal shallow understanding.Open details

Details

Subscribing short-lived objects to singleton events.
Using anonymous handlers that cannot be removed easily.
Assuming the subscriber disappears when it leaves the UI or request.
Adding weak references without understanding ownership.

Red flagsCandidateRed flagsSignals that an interviewer may treat as risky.Open details

Details

Cannot describe the publisher-to-subscriber reference.
Does not ask which object lives longer.
Only suggests increasing memory limits.
Ignores lambdas and anonymous event handlers.

Follow-up prompts

Quick interviewer prompts inside this drill.

Follow-up questionsCandidateFollow-up questionsLikely next questions after the main answer.Open details

Details

Treat these as the next interviewer turn. Answer one out loud, then save the gap in your practice notes.

Who holds the reference after an event subscription?
When would Dispose unsubscribe?
Why can anonymous handlers be risky?
How would a heap dump reveal this leak?

Continue the section

Next in GC and Lifetime: Singleton Capture

Step 8 of 8

Why is it risky for a singleton to capture scoped services or request data?

A component subscribes to an event and memory grows over time. How can the subscription cause a leak?

Practice focus

Answer path

Why this question matters

Review signals

Follow-up prompts

Continue the section

A component subscribes to an event and memory grows over time. How can the subscription cause a leak?

Practice focus

Answer path

Why this question matters

Review signals

Follow-up prompts

Continue the section