Menu

Feature Request: Date Range Search

2010-05-04
2013-02-02
  • Pete hoffswell

    Pete hoffswell - 2010-05-04

    We have a standard workstation naming convention that allows us to identify a group of computers in a computer lab.  With this information in hand, we can actually use NetDB to give us a utilization rate for a specific location and room.

    One thing we cannot do is look at utilization for a date range.  Current (NetDB v1.7)search only allows for a specified number of days back.

    It would be handy for our utilization research, and probably other auditing functions, if you could search by date and date range:

    Query: SITEXROOMY
    Report Type: Host Name
    Start Date: 4/1/2010 (Leave blank to count backwards from today)
    Number of Days: 1

    For our naming convention, this would return all computers at Site X room Y that showed up on the network on 4/1/2010.

    Good stuff.  If it makes sense for the project, I hope you will consider adding it.

    Thanks!

     
  • Jonathan Yantis

    Jonathan Yantis - 2010-05-04

    Hey, I'm going to have to think hard about date ranges based on the way the program operates and stores date, but it may be possible.  I'm a little worried about the accuracy of the results though, since historically you can't tell if that IP was in use at that exact time by a mac address, only when it was first seen and last seen.

    One way to capture this data like this is to look at the update-statistics.sh script in extras.  You could start graphing or recording data in real-time with something like this in cron:

    /opt/netdb/netdb.pl -n sitexroomy -h 1 -c  | wc -l > /opt/netdb/data/sitxroomy.txt

    And then you could start graphing that data with MRTG or some other tool like cacti.  Over the long run, that data will be much more accurate.  If you have a machine that is in the room one month, gone the next, and then back again the third month, looking at it historically, it looks like it never left.  Recording that data in real time with rrd or something similar would eliminate those anomalies.  If I changed the program to record every flap of every device, the database would unfortunately explode with data. 

    How does this sound?

    By the way, I'm working on the 4-digit mac query which should be there in the next major release.

    Jonathan

     

Log in to post a comment.