A nice thing in Celery Flower is viewing the `args, kwargs`, whereas Hatchet operates on JSON request/response bodies, so some early users have mentioned that it's hard to get visibility into the exact typing/serialization that's happening. Something for us to work on.