Problems in the QEE Hlt2 Monitoring histograms
FYI @lugrazet
cc @abertoli @mvesteri @poluekt
I've recently taken a look at all the monitoring histograms in /eos/lhcb/wg/rta/WP3/HLT2_Monitoring/HLT2-run268205-268264.root
.
I've drawn them and put them all in 1 .pdf here: qee_monitoring_plots.pdf.
The issues I found (not necessarily all the issues that are there):
-
ALPsToGammaGamma
: strangely bumpy mass distribution given the statistics FYI @odurmus @tmombach, -
Di{Light,TopoTag,SVTag}{10,15,20,25,30,35}GeV
: mass, ipchi2 and vchi2 histograms are mostly unfilled/entries are in overflow/underflow etc. FYI @dzuliani @ncooke @ngrieser - Some funky shapes and lack of quarkonia mass peaks in the
DiMuonNoIP
lines. Will elaborate on that in the discussion. FYI @kaaricha, -
{Jpsi,Upsilon}ToMuMu
: no discernible mass peaks. A couple of strange eta distributions. FYI @lugrazet -
WZRare
: vchi2, ipchi2 and mass sometimes don't make sense. A couple of the lines have combinatorics issues. Will elaborate on that in the discussion. FYI @hyin. - All lines in
high_mass_{dimuon,dielec}
e.g.ZToMuMu
,ZToEE
,DiMuonSameSign
etc: mass histograms incorrectly filled/don't make sense e.g. events all in under/overflow; axes limits wrong way around etc. FYI @jideng @hyin - No histograms present for any of the
SingleHighPt{Muon,Electron}
lines. FYI me, @lugrazet - No histograms present for the
DiElectron
lines underdielectron_persist_photons
. FYI @johndan
For those interested, the monitoring is defined in Moore/Hlt/Moore/python/Moore/monitoring.py
. The answers to some of these issues may be found in the configuration/assumptions made there.
Very relevant talk from Marian Stahl: https://indico.cern.ch/event/1234789/#7-hlt-line-monitoring
Outstanding Concerns
The problem related to N_CANDIDATES being incorrectly filled, mentioned several times in this issue, is being addressed in !2798 (merged).
The electron-related issues are still present in the default monitoring. see: #653 (comment 7319635)
The high_mass_di{muon, elec} is still not returning histograms. see: #653 (comment 7077762)
UPDATE AS OF 30-Apr-24: These three issues have all been handled... think once !3352 (merged) goes in it can be closed