Show "Bottom 10"

That solution fails for SSL certificate expiration days remaining, unless you like seeing negative days.  We found a workaround by using a max of 90 days to see only those expiring in that period, but bottom 10 (or N) would be very nice to have as a general tool.

I pretty much lost hope of this ever happening, but perhaps one day!  I have a datasource I am still trying to finish that measures the "% used lifetime" of a certificate so that top-N can be meaningful since that is all we have.

Hi - We agree that displaying bottom 10 can be valuable, and we typically recommend that you do this by using a virtual datapoint to inverse the data & then limiting to top 10.  For example, with CPU busy percent you can create a virtual datapoint in the widget equal to 100 - CPUBusyPercent & limit that to top 10 - that gives you the top 10 least busy CPUs (or the bottom 10).  

That makes sense - we'll look into adding a bottom 10 (or N) option. 

Having a Bottom 10 would help with finding devices that don't stay up for very long.  Or, in my case, trying to find the collectors that are restarting the most before the 8 hour window.  For now, I did make a Virtual datapoint to subtract the Collector Uptime from 480 for the top 10, but then people reading it would really need to "know" that collectors are restarted every 8 hours.

Being able to just show the Bottom 10 that have the smallest Uptime would be easier to understand by the ones that will be looking at the graphs.

On 1/27/2017 at 2:37 PM, mnagel said:

That solution fails for SSL certificate expiration days remaining, unless you like seeing negative days.  We found a workaround by using a max of 90 days to see only those expiring in that period, but bottom 10 (or N) would be very nice to have as a general tool.

Mark

We're also running into the SSL cert expiring issue & would be beneficial to have a "Bottom 10" so we can see which ones are the more urgent to resolve.

My new SSL check is virtually done, just need to DS-ize it.  The code works for any site, including SNI.  This means no AD since you can't discover what URLs a host will respond to, but it will work on multiple URL instances in the DS when I am done.

[mnagel@colby ~]$ groovy getCertv2.groovy https://www.google.com
Response Code : 200
Cipher Suite : TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256

         Subject: CN=www.google.com, O=Google Inc, L=Mountain View, ST=California, C=US
          Issuer: CN=Google Internet Authority G2, O=Google Inc, C=US
        Used Days: 10
   Remaining Days: 74
Lifetime Consumed: 12.0%

Alerts will still generally be tied to remaining days, but graphs will be able to show "percent lifetime consumed" and alerts could be tied to that instead if preferred.

Our use case for this is that we have a large VMware environment with thousands of VMs and hundreds of datastores.  Being able to see the unused and underused resources helps us identify left-over items to prune or consolidate.

Also "most recently rebooted" systems, least used interfaces, etc.  There are many legit reasons for inverting and it seems like a fairly simple change (to me anyway).  Hopefully soon....

Thanks,
Mark

Bottom 10 is coming soon (estimated in the next few months).