Improve error handling

[sam-grant]

It's a bit sloppy at present.

Example: everything falls over but the exception is not propagated up.

[Analyse] ❌ Error during analysis execution: 'CutManager' object has no attribute 'calculate_cut_stats'
[CosmicProcessor] ❌ Error processing /exp/mu2e/data/users/sgrant/mu2e-cosmic/data/nts.mu2e.CosmicCRYSignalAllOffSpillTriggered-LH.MDC2020as_best_v1_3_v06_03_00.001202_00050440.root: 'CutManager' object has no attribute 'calculate_cut_stats'
[pyprocess] ❌ Error processing /exp/mu2e/data/users/sgrant/mu2e-cosmic/data/nts.mu2e.CosmicCRYSignalAllOffSpillTriggered-LH.MDC2020as_best_v1_3_v06_03_00.001202_00050440.root:
'CutManager' object has no attribute 'calculate_cut_stats'
[CosmicProcessor] ✅ Analysis complete
[run] ❌ Analysis failed: name 'info' is not defined

I think top level parent processes should sys.exit(1) on a critical error, and then worker processes should just raise and propagate the exception up. There may be instances where return None is good, need to think about it.