I'd like to use TabPy in my enterprise, but I have concerns about the impact it would have on dashboard rendering. Can anyone address these questions?:
Are execution requests to TabPy single-threaded? That is, if I have 10 views rendering with Python code, will the be handled serially or in parallel? I'm curious what the impact would be to viz rendering times if TabPy gets backed-up.
Can I run a TabPy farm behind a load balancer? This could be a work-around if TabPy is single-threaded; the load balancer would round-robin requests to TabPy servers based on their availability.
Does TabPy have to run as admin? I worry about someone slipping in a "os.system('shutdown')" command (or worse).
How do I automate restarting the TabPy instance? I'd like to set up some automation where an instance gets restarted after each script is run, to prevent anyone from maliciously investigating the previous script's data space.
Any plans to re-work TabPy as a Windows service? The startup folder hack doesn't really work with monitoring and automated restarts.