...

• Every cnf needs a different platform (for collection code)
  • since platform corresponds to the primary readout group, this ensure that the primary readout group is different for different users, but also have to make sure other lower-rate readout groups are not used by different people at the same time
• All cnf files should use the same hutch supervisor xpm (e.g. xpm2 for TMO)
• Timing node can be shared, but each cnf should use a different fiber lane
• Different cnf files should not startup the drp executables that others want to use: can create errors since if file descriptors are in-use by another user's executables
• Be careful with configdb setting conflicts: there is a configuration database per hutch/user defined in the cnf file (e.g. tmo/tmoopr) and there is also a prod and a dev configdb. 

daqPipes and daqStats

Useful commands from Ric for looking at the history of DAQ Prometheus. Note that daqStats shows raw counts, while daqPipes shows percentages and also highlight (in red) detectors that are backed up (full buffer). daqPipes is more useful for understanding where bottlenecks are. 

...