Austin Culbertson

  • Posts

  • Joined

  • Last visited


0 Neutral

About Austin Culbertson

  • Rank
  1. Hrm... It sees to me that the only way I could calculate percentiles is if I ran multiple reports for the time frame I want (i.e. day to day, for 7 days), and then calculated 95th percentile from that? Unless there is some other fancy math that I'm missing that can be applied to the 'in/max/average/change' values...?
  2. Sorry, I don't follow - I do not see '95th percentiles' being present in the resource metric reports?
  3. Got it - That matches my understanding. Appreciate it. I'll key a feature request for having percentile calculations available in a report - luckily, I just found out that this _isn't_ a requirement for my current solution, so I'lll be happy with just the visualization at this point
  4. Am I right in assuming that this will only allow users to see a _visualization_ of these percentiles? This pretty strongly defeats the purpose of the use case for percentiles, which are most commonly needed for _text_ based reporting.
  5. Do there presently exist any methods for generating advance resource metric trends, by chance? I'm thinking something along the lines of '95/99th' percentiles and the likes, in the form of a report. If not, what's the next best thing?
  6. Greatly appreciated! I'll take a look at this today and pick it apart.
  7. Unfortunately, Service Insights is not an option for us, as we would need to configure on the order of 1000+ Services, each with maybe only 1-4 instances for data being aggregated. Not really an option for us, unfortunately (I've spent a lot of time banging my head against this particular wall ) Do you know what library plays nicely in this case? Is GPars the right one? I'm just trying to find something that will work -- would be best if it's something that works natively with existing libraries (though it shouldn't really be a problem if not). [edit] Herp, I forgot to read the Stack Overflow -- sounds like maybe going straight Java might be best. I'll see if I can work with this -- Thank you!
  8. Has anyone tried to use any type of code-level multi-threading within a DataSource, by chance? Is this even possible? My use case is for querying 16 different hosts for SNMP information and then aggregating the data based upon certain properties - Think "Combine traffic amongst all TenGigE0/0/0/10 interfaces that have the description 'P2P connection', then all TenGigE0/0/0/10 interfaces with description 'P2P interfaces for Zayo'" -- This is not my _actual_ use case, but an approximation. At present, collection times are hovering around 110 seconds to run through all the operations we need, which is getting a little close to the 2 minute mark for my tastes -- I'd prefer not to have to develop a whole suite of tooling to take advantage of Push metrics if possible, so I'm trying to see what I can do to optimize my groovy script, and it seems like threading is the next logical answer for this; that being said, I've not done anything multithreading before, so there's that I've tried using GPars, but it doesn't seem like it's installed by default (so I installed it as a custom library), and I'm having a little trouble getting it to work (which might be because this might not be a thing that I can do, for myriad reasons?).
  9. Does anyone know if it's possible to invert the devices that make it to the 'top' spot from an overview perspective? As I understand it, it's Top 10 that make it to these overview graphs -- what if I want to see the 'Bottom 10' on a given Overview Graph? I imagine there is probably some complex datapoint that I can use to invert my value, but this would add to the complexity of the datasource, and I'd like to keep it simple if possible.