-
Notifications
You must be signed in to change notification settings - Fork 1.6k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Running tasks creates linear memory growth #14329
Comments
If the task engine is updated to explicitly drop in memory results, the memory usage decreases and stays flat across invocations. 10 task runs:
100 task runs:
This suggests a more sustainable solution to avoiding caching results in memory will resolve this issue. |
I am also dealing with memory issues in Prefect, and it seems similar to the one you have described. |
Hi @frankvp11 - this issue for Prefect 3.0. If you're on 3.0 please follow along here. If you're using Prefect 2, you may find #12668 a more relevant issue. |
The following script is an MRE to demonstrate linear memory growth when running many tasks in a flow:
Running the flow with 100 tasks produces the following profile:
Running the flow with 1000 tasks produces the following profile:
We'd expect the memory usage to be consistent across runs since the flow isn't holding onto any return values, which suggests that we are holding onto results somewhere that we shouldn't be.
The text was updated successfully, but these errors were encountered: