Re: Inspect Samples of CSS shows inconsistent SEVR

["Kasemir, Kay" <kasemirk@ornl.gov>]

Hi:

The CSS data browser, its sample inspector etc. _show_
the severity. They don't determine it.
The "Channel Archiver" source in your example indicates that
your samples are provided by the Channel Archiver's network data server.

When that network data server is asked for optimized data,
it performs interpolation via the "plot binning" described
on page 14 of http://ics-web.sns.ornl.gov/kasemir/archiver/manual.pdf.
The values you see are the initial, minimum, maximum, final values within
each plot bin, though the protocol doesn't really indicate which one is
which.
The status/severity is not really handled for those min/max samples.
The point of the plot binning was to have something that looks 99% the same
when plotted. It's not meant for introspecting each sample.

Thanks,
Kay





From:  Burkhard Kolb <b.kolb@gsi.de>
Date:  Wed, 4 Apr 2012 07:51:29 -0400
To:  "tech-talk@aps.anl.gov" <tech-talk@aps.anl.gov>
Subject:  Inspect Samples of CSS shows inconsistent SEVR


Hi,

when looking at archived data with the CSS trend plot "Inspect Samples"
tool we
found some surprises in the archived data:
It looks like the interpolated samples show the wrong Severity and Status.
When I export the raw data from the archiver all value/status/severity are
ok (in this case the LOW limit was set at 20).
See screen shot of CSS(3.1.0.20120201):