Fix the memory leak of Google STT when no audio input.#5591
Merged
longcw merged 4 commits intolivekit:mainfrom May 7, 2026
Merged
Fix the memory leak of Google STT when no audio input.#5591longcw merged 4 commits intolivekit:mainfrom
longcw merged 4 commits intolivekit:mainfrom
Conversation
When the RTC session is idle without any audio input, Google STT re-create the client object and re-connect periodically. However some old resources actually remain as they're not cancelled properly, causing memory leak. This PR adds the clean-up of such scenarios. With this change, I confirmed my experience of the memory leak is gone. Fixes livekit#5590
|
|
zelidrag-arbo
approved these changes
Apr 29, 2026
longcw
reviewed
May 6, 2026
Contributor
longcw
left a comment
longcw
left a comment
There was a problem hiding this comment.
thanks for the pr! overall looks good to me, my only concern is about the await asyncio.sleep(5.0) after timeout
| return | ||
| try: | ||
| frame = frame_task.result() | ||
| except StopAsyncIteration: |
Contributor
There was a problem hiding this comment.
use ChanClosed if using the input_ch.recv on above
| else: | ||
| # No audio has been pushed — back off so we don't tight-loop | ||
| # against Google's server-side 5-minute idle timeout. | ||
| await asyncio.sleep(5.0) |
Contributor
There was a problem hiding this comment.
should we await sleep here? what if user audio comes in during the sleep?
Contributor
Author
There was a problem hiding this comment.
Indeed, didn't think much about such cases; removed.
Co-authored-by: Long Chen <longch1024@gmail.com>
![devin-ai-integration[bot]](https://avatars.githubusercontent.com/in/811515?s=60&v=4){kind=small}
This comment was marked as resolved.
This comment was marked as resolved.
Sorry, something went wrong.
jmuk
commented
May 6, 2026
Contributor
Author
jmuk
left a comment
jmuk
left a comment
There was a problem hiding this comment.
Thanks for the reviews!
| else: | ||
| # No audio has been pushed — back off so we don't tight-loop | ||
| # against Google's server-side 5-minute idle timeout. | ||
| await asyncio.sleep(5.0) |
Contributor
Author
There was a problem hiding this comment.
Indeed, didn't think much about such cases; removed.
| return | ||
| try: | ||
| frame = frame_task.result() | ||
| except StopAsyncIteration: |
longcw
approved these changes
May 7, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
When the RTC session is idle without any audio input, Google STT re-create the client object and re-connect periodically. However some old resources actually remain as they're not cancelled properly, causing memory leak.
This PR adds the clean-up of such scenarios. With this change, I confirmed my experience of the memory leak is gone.
Fixes #5590