Skip to content

Monitor ROS requests in TrigCostAnalysis

Aleksandra Poreba requested to merge aporeba/athena:ros2rob-file into master

For each data request MonitorROS creates a counter that will process ROS request. To resolve which ROS corresponds to which ROBs (data request consists only of a list of ROB ids) a map was added (ROSToROB.py). The information was retrieved via script ros2rob_from_partition.py.

If ROS request serves >= 1 ROBs the CounterROS is created and data is saved for all ROBs that are in current request (data.rosCollection()) and correspond to the current counter ROS (data.rosToRobMap().at(rosName)).

Data request information is saved to histograms:

  • Request_perEvent - number of ROB requests per event
  • NetworkRequest_perEvent - number of ROB requests per event fetched over the network
  • CachedROBSize_perEvent - size of cached ROB requests
  • NetworkROBSize_perEvent - size of ROB requests fetched over the network
  • ROSTime_perCall - elapsed time of processing request
  • ROBStatus_perCall - history of ROB request, the last bin is if ROB header is ok

Example histograms for ROS ROS-SCT-B-00:

status requests

Tagging @mark @tamartin

Edited by Aleksandra Poreba

Merge request reports