Time-dependent histogram for monitoring SciFi hits

One thing in the failures which is a bit worrying (but unrelated to this MR) is that sometimes the file to run over cannot be found (something I had observed privately before).

Ok, I'm passing it to @jonrob for merge.

which test(s) does this monitor run in ?

@decianm ping - Are there any QMT tests that exercise this monitor ?

Anything that runs the monitoring runs this monitor, e.g. hlt2_light_reco_pr_kf_with_mcchecking

So you expect to need a bunch of ref updates then yes ?

I did not see any test that needed ref updates (I see that the monitor is running in the test, but it does not produce an output for a reference).

Turns out this did need a ref update. See the Rec today ...

https://lhcb-nightlies.web.cern.ch/nightly/lhcb-master/2097/

!3514 (merged)

assigned to @mzdybal

assigned to @jonrob

added FT Monitoring labels

@decianm I think if you want to ensure the histograms are reset on a run change, you probably need to do this yourself in the monitor, by e.g. detecting when the run changes and do a manual reset.

One other thing to bear in mind is when running at the pit, run changes as seen by any online task is not a black and white transition. Due to the way the events get sent to the monitoring from HLT1 you can get for a short period a mixture of old and new events coming. You might also need to think about how to deal with this as well.

I think if you want to ensure the histograms are reset on a run change

What do you mean? I don't think any monitor resets histograms when there is a run change (?)

you can get for a short period a mixture of old and new events coming

That should then affect only the last or the first bin, and I assume the overlap only exists for some events, and not for seconds in a row. So I simply assume that this effect will be small.

mentioned in issue Moore#623 (closed)

assigned to @jagoodin and unassigned @mzdybal

added 66 commits

de777024...fedc4020 - 59 commits from branch master
663c1ddb - first attempt at time-dependent histogram for monitoring SciFi hits
8f40752b - Fix wrong use of 'second'
34293dd0 - fix wrong understanding of histogram filling
d1891f3c - fix for calculation of number of hits
b9e52adb - Fixed formatting
bbfa7cee - split histograms in top and bottom
f0a8775c - Fixed formatting

Compare with previous version

resolved all threads

merged

mentioned in commit 04416cac

mentioned in issue Moore#627 (closed)

Time-dependent histogram for monitoring SciFi hits

Merge request reports

Activity