High memory usage on asset with lots of JSON GET requests

Anthias
1

I have been looking into an issue with my in house dashboard I built and Anthias. Short of the long there is a single asset webpage that makes lots of JSON GET request’s that vary from every 10 seconds to every 15 min depending on the data they return. Hardware is a Pi 4B 8 GB RAM.

I have noticed a slow but steady uptick in memory usage when said asset is displayed until I get the black screen (BlackSOD haha). The web UI remain accessible the whole time.

If I load a simple page such as example .com I do not see this issue and the memory stays steady so I assume it has something to do with all GET’s the page does.

I plan to install Anthias on top of the PI OS this week vs the Balena images so I can grab logs.

For now has anyone seen this issue before and was it an issue in your JS, something with how Anthias does the GC or something else? Maybe I am pushing Anthias beyond it’s limits.

I think my issue is related to these posts:

2

Are you on the latest release (from master)?

Also, web pages often leak memory. The longer you leave them up, the more likely you are to hit it. A quick workaround is to add an image in the playlist rotation to clear that. Alternatively, you can try the new ‘reload’ feature, but it isn’t as bullet proof.