This discussion has been locked. The information referenced herein may be inaccurate due to age, software updates, or external references.
You can no longer post new replies to this discussion. If you have a similar question you can start a new discussion in this forum.

Orion Exact-Timeframe Polling

Perhaps the subject isn't descriptive enough, but I've run into a couple of data polls where there has been a gap in some of our data, which often shows up as a blank space on a graph.  Looking into this further, I saw data points reported to be 5 minutes apart, with the occasional data point 6 minutes out. Having done some custom monitor stuff with shell scripts in the past (a true horror, really) I think I know where this is coming from.

It seems as if Orion starts the poll, completes the poll, submits the data, then waits the defined interval before the next poll.  However, since polls are not necessarily instant events, the poll cycle extends to the poll_time+interval.  For example, if it takes 5 seconds to do a poll, the interval is really 5 minutes, 5 seconds.

My suggestion is to take the start time before the poll, do the poll, then wait for the remainder of the time before the next poll.  The change might require some testing, but shouldn't be too difficult to implement.

  • If you don't mind me asking, why do you need such precise polling intervals? I'm not trying to create trouble, just curious.

  • The gaps in the graph are tiresome to explain to management.  :) 

  • In that case can I suggest that you modify your feature request? Instead of asking for super high precision polling ask that the graphs be displayed as area charts instead of vertical bar charts. This makes more sense anyways and should allow your graphs to flow freely without gaps. This seems to be the direction they are taking the next version of NTA anyway, so it's not like they'll be reinventing the wheel. Just a thought.

  • The chart is largely what I need ... save the issue with the gap.  I don't think that it's unreasonable that a 5-minute poll be every five minutes, on the interval.  At least any interval observations I did in my college labs would have been marked as incorrect had I followed an additive interval.

    I don't think it's unreasonable to have this minor issue corrected.
     

  • This is major problem for us right now. I rebuilt the database Friday; Saturday I checked the data, everything looked good (with a few blanks). I showed up to work Monday and we are back to spotty data!!! These charts are one of the major requirements for purchasing this product.What you have described here makes perfect since, and would explain why as time goes on I see more and more gaps between data collection.

    I am currently looking for a fix!!