our platform (tutor version 15.1.0, local deployment on a 16 GB Memory / 320 GB Disk, ~7k users with 10k unique enrollments over 10 courses) is periodically running into the issues described here and here. In short, as the platform runs, the reddit files at $(tutor config printroot)/data/redis become large. After a while, they get to about ~10GB at which point celery tasks stop working. I can manually delete the files and restart reddit and the cms-worker images (as described in the topics linked above) and then everything works again. The issue is that at first it took almost a year to get to that point, but as our users grew it now it barely takes a month.
As suggested here, there should be a way to configure redis to “trigger a rewrite at the right time”, however I have not found any documentation on how to do that within the tutor setup.
I understand that directly editing $(tutor config printroot)/env/apps/reddit/reddit.conf is not the way to go since the env folder will be regenerated every time I do tutor config save.
However, I also understand that I cannot use a plugin via openedx-lms-common-settings with the redis-conf template patch unless I upgrade tutor to a version post your linked commit. I have instead tried a custom plugin, i.e.:
otherwise, it just occured to me, may I use tutor >= 17.0.3 with openedX olive? this way I could use the redis config fix without having to upgrade openedX.