cmsdaq / DAQExpert

New expert system processing data model produced by DAQAggregator
1 stars 2 forks source link

FED deadtime for deadtime analysis vs current warning #155

Closed gladky closed 6 years ago

gladky commented 6 years ago

There is a LM to be introduced that check whether FED DT > 2% and FED BP < 2%. The thresholds will be parametrized from a configuration file. We have currently the FEDDeadtime LM that warns when individual FED deadtime exceeds given threshold given in conf with key:

expert.logic.deadtime.threshold.fed

Few questions:

  1. Should the deadtime thresholds for this LM and for existing FEDDeadtime be based on the same configuration entry (expert.logic.deadtime.threshold.fed) or independent? Currently the warning threshold is 2% similarly to what was described for deadtime analysis.
  2. Should the backpressure threshold for this LM and for exisiting BackpressureFromEventBuilding, BackpressureFromFerol and BackpressureFromHlt be based on the same configuration entry (expert.logic.deadtimeanalysis.fed.backpressure.threshold) or indepenedent?
  3. Should the deadtime analysis be performed regardless the fact that there is 100% deadtime or the rate is zero? Are there any conditions when deadtime analysis is pointless to do and can be filtered? E.g. only during running?
gladky commented 6 years ago

Assuming answers:

  1. yes
  2. yes
  3. no

and closing.