snoplus/minard

Minard - SNO+ Monitoring

minard's Issues

All of the Data Quality pages are extremely slow

404 not found from PMT Noise page

When clicking ratdb upload log file I get a 404 Not Found.

Sparkline graph doesn't show time

Currently, the sparkline graph for running the Snoplus detector shows the date in the top right corner when hovering the mouse over the graph according to its point on the x-axis, but it doesn't show the time. It would be very useful if it did, as we could tell the exact time events happened instead of approximating, and would make things easier to analyse.

Detector State 0 page shows incorrect enabled triggers

The detector state /0 page shows the wrong set of enabled triggers (and the thresholds for those, rather than the correct set of enabled triggers). I've attached screenshots.

Pop-up box with incorrect info on detector state page

On http://snopl.us/monitoring/state/ I get a box when the mouse hovers over the thresholds saying they are not masked in, when they are enabled (and orca says they are masked in)

detector state data is incorrect for missing slots

The problem is that the function fetch_from_table_with_key assumes that if key is None, it should find the maximum key in the table, but if a card is missing then that column will be NULL which gets translated into a None object in python.

One solution which would simultaneously fix this problem and speed up the database queries would be to use the detector_state view in the database. Something like:

SELECT * FROM detector_state WHERE run = [run];

TUBII attentuated channels should be in database

Replacing this:
https://github.com/snoplus/minard/blob/master/minard/detector_state.py#L294-L323

@azummo

500 Internal Server Error

When accessing https://snopl.us/monitoring/deck_activity, I get the message:

500 Internal Server Error
There was an internal error on minard. You can report issues on the github page here https://github.com/snoplus/minard/issues.

When a card is pulled or not in the polling slot mask the base currents get misreported

All the cards after the missing one are shifted one and the one immediately after the missing slot has values of -127.

See all baselines

Would be handy to be able to view of all the baselines simultaneously.

Shifter Information page crash

When filling the phone number in the Shifter Information page, it complains that the number is not the right length (the check should be country-dependent).
Then when trying to submit page again without the number, it crashed. ("500 Internal Server Error").
And stayed crashed in following days.

RAT page returns Internal Server Error

We've stopped running the ANALYSIS job on the nearline as of run 115111 so that might be related.

NHit monitor should have drop down for selecting different types of triggers

HV status in detector state

Could there be another crate-view section that shows the HV status of the crates? So an user off-site can see which crates are on HV easily?

CAEN channel offsets not interpreted correctly

Despite the data saying channel 2 has a different offset the display does not show any difference. Also most of the values should read 0. They instead read -1V.

Add NONE type triggers to minard stream

They can appear for some types of retrigger events.

timelines x and y axis should be re-scaleable.

Stacked trigger-type time-series

As was done in Snostream, we could display trigger-type over time by using a stacked histogram, see screenshot below. I am not sure how to interpret the x-axis there though. 🌳

Small detail: time-scale in stream vs. trigger detail page

The stream time-scale is on Sudbury time, but when you click on a stream-bar for more detail, the time-scale of the time-series is in local time. I think it should also be in Sudbury time?

Detector state page should show pulser rate

Maybe its there, I couldn't find it.

trigger thresholds displayed incorrectly for current state

The current trigger thresholds do not match the values reported for the last run. The issue appears to be due to an incorrect trigger scan being fetched since the timestamp for the current run is not updated.

Reloading the stream page does not force an update of graph. [low priority]

Using an auto-reload extension or manually reloading while on a stream page (say 100M Baseline) does not update the graph. Leaving the page and returning will update the graph.

periodic shuffling of events between bins

As was seen yesterday, the stream page seems to sometimes blink and then shove the missed events into the next time window. Produces nonphysical spikes in the time series. This appears to be happening with a regular periodicity of 3.5 minutes currently - perhaps that suggests where to look for the problem. See attached image.

Zoom/enhance baseline drift

It'd be good if the baseline drifts were 0 centered over the given time window and perhaps given in mV instead of volts.

Time scale in trigger details page

Add the time-scale drop down also on the trigger-specific page. This way, people can change the time-scale directly on this page, without needing to go back to the stream page.

pmt noise rate should be plotted versus time

Currently it's very difficult to get an overall feel for the PMT noise rate since the plot is displayed versus run number.

This could be solved by plotting versus time instead of run number.

It might also be cool to display a native graph on the page instead of a ROOT plot so you can select different time ranges, etc.

@edcallaghan

DQ pass/fail checks should give some short explanation of what they check

That way a shifter looking at the data quality pages can figure out why a particular run might have failed some check and (hopefully) correct it.