High memory consumption after opening a file resulting in IDE issues

[tmirza]

Question

Hi,
I have deployed my Theia IDE browser application using a Docker container running on an EC2 instance in an AWS chamber. I have noticed that if I open a large-sized file, I get a dialog about file size, which is fine, but clicking on Yes, Theia backend tries to load the contents of this file and stream to the browser. This results in a sudden spike in memory consumption of my Theia IDE container, and even if I close the tab, this memory is never reclaimed from the node process.
For an EC2 instance with 8GB memory, I see the following in Docker stats for my Theia IDE container after a while.

ea665adc1896   theia-ide      9.82%     6.235GiB / 7.499GiB   83.14%    62.7MB / 28.5GB   4.65GB / 40.3MB   37

Is it a known limitation of Theia? Is there any way to reclaim this memory without restarting the Theia IDE container? This huge memory consumption significantly slows down all IDE operations and bogs down the overall IDE performance.

cc: @akosyakov

[github-actions[bot]]

Hi @tmirza, thanks for starting this discussion! 👋

The Theia community will take a look soon. In the meantime, you might find helpful information in:

• 📖 Theia Documentation
• ❓ FAQ
• 💬 Previous Discussions

---

💙 Eclipse Theia is built and maintained by a community of contributors and sponsors. If Theia is valuable to your work, consider sponsoring the project. For professional support, training, or consulting services, learn more about available options.

[msujew]

Hey @tmirza,

first off, please don't randomly ping contributors to the project. Anton hasn't contributed here for 5 years now, so you probably won't receive an answer from him anyway ;)

Secondly, how large is the file you're trying to stream? It's not usual for node.js processes to keep a large amount of memory, even after reclaiming it during GC internally. Looking through our issues, I haven't seen anything about this problem so far.

[tmirza]

@msujew ,
My apologies for the direct ping.

Regarding the issue, as per my understanding, every attempt to open a binary file results in Theia backend reading the content of that file and I see same amount of jump in node process memory consumption. In above scenario, there were 5 or 6 files of 1.5G that I clicked and then Theia tried to open them and on every file opening, it resulted in memory usage spike and after opening two or three files, my browser tab crashes and I have to reopen my IDE in new tab. I think that is because of backend streaming the file and resulting into OOM. Even after browser tab crash, when I look at node process memory, it still stays the same.

Thanks.

[msujew]

I see - if the browser crashes during the streaming, I can see the potential for the backend to keep some dangling pointer to the data somewhere. Feel free to open an issue about that with some clear reproduction steps.