Memory issues (long running instances)


I've a lot of Filebeat instances (v7.2.1) running on separate containers and noticed that longer an instance is running, the more memory (RSS) it uses. While a restart helps and the memory usage drops (e.g., from 380MB to 15MB after ~80 days of continuous running) I don't see that as a viable option.
Is there anything I missed during searching for various FB memory leak threads?
AFAIK, there isn't any option to limit the memory usage of the process right?
Is there any workaround other than restarting the process? If not, I'm concerned that during the restart and the initial reading of the input configurations (5 minutes for me and it's set to that for good reason) it might miss files that were rotated (close_renamed).

Thank you!