Should we expect monitor events for non-VAL fields on SEVR change?

[André Favoto via Tech-talk <tech-talk at aps.anl.gov>]

Hi all,

Replicating a matrix channel question here:

I have noticed a behavior that I'm not sure if it is like this by design or not.
When a record SEVR changes, we only get monitor events for the VAL field. This happens either on rising or falling edge.

Since fields do inherit the alarm state, I'd assume they would react to it being changed as well. Example with motor-record (but also applicable to others):

$ pvmonitor YMIR-BmScn:MC-LinY-01:Mtr.ACCS YMIR-BmScn:MC-LinY-01:Mtr.RBV YMIR-BmScn:MC-LinY-01:Mtr.DESC YMIR-BmScn:MC-LinY-01:Mtr.HIHI
YMIR-BmScn:MC-LinY-01:Mtr.ACCS 2026-10-05 15:00:00.735  8 MAJOR DEVICE STATE
YMIR-BmScn:MC-LinY-01:Mtr.RBV  2026-10-05 15:00:00.975  11 MAJOR DEVICE STATE
YMIR-BmScn:MC-LinY-01:Mtr.DESC 2026-10-05 15:00:00.975  BeamScanY MAJOR DEVICE STATE
YMIR-BmScn:MC-LinY-01:Mtr.HIHI 2026-10-05 15:00:00.975  0 MAJOR DEVICE STATE       ## I reset the error here
YMIR-BmScn:MC-LinY-01:Mtr.RBV  2026-10-05 15:00:10.926  11                                         ## only .RBV (or .VAL) reacts to it

If I close the monitor and open another one (or do a pvget), then the alarms are shown as gone, which to me indicates it's a monitor-event-posting thing.

It would be natural to me that either a) non-VAL fields do not inherit alarm states at all or b) if they do, they react when SEVR change.

PS: This question came from using OPIs showing multiple record fields (e.g. motor record):

The alarm border on the field widgets never clears unless we reopen the screen, so alarm sensitivity needs to be disabled.

If that's the expected behaviour I'm happy to contribute with a small note about it on the docs.

Best Regards,

André Favoto