How to debug sporadic memory bugs in a complex Python data pipeline?

[NoraEG]

I'm working on a fairly complex Python data pipeline that's started throwing a sporadic, non-deterministic error related to memory management, and my usual print statement debugging isn't cutting it. I know I should be using a proper debugger or profiler but I've always found the setup clunky. For intermediate developers who have moved beyond basic debugging, what tools or workflows do you recommend for isolating tricky issues in larger codebases, especially ones that involve concurrency or external libraries where stack traces aren't immediately clear?