@Gargron @bclindner we upgraded from 3.1 to 3.3 and that’s when it apparently started. I was keeping an eye on it before that and memory use appeared stable, though our infrastructure was not that stable so frequent (1.5/week) reboots may have obscured any problems. Happy to provide more context!

(Also: I have committed to moving off Docker, I just need to find some time to do it 😬 Thank you for peer pressuring me all these years.)