We have about 1364 hosts that I was asked to generate a real-time report on the number of clients that have connectivity issues and are down. On the dashboard, I can create a custom tile that will show the connectivity chart but only is able to get 15 devices. Instead of the chart would be great to have a numeric value indicating how many hosts have lost connection, and highlight the hosts when opening up the tile
Answer by Kris Z. ·
I'm not sure it would be possible to do it on a single tile. Bot for sure it's doable on two tiles that can be shown side by side:
use row counter tile to present a number of rows in a table that has dimension of server IP and metric of Failures (TCP) filtered >0
use metric tile to present value of the Unique users metric, while filtering by Failures (TCP) >0
First tile will show number of servers that returned any TCP failures. Second tile will show number of users who experienced TCP failures. Then add drill downs to lists of hosts, users, etc as needed. Perhaps straight to Operations Explorer or Error Explorer to make things simpler.
Answer by Frans S. ·
I am following this thread with interest, and like the good answer by Kris.
But I am also particularly interested in the Host perspective view. And I think that is what Sam also wants.
A list of (unique) hosts that have availability (connectivy) issues, and failures. And the number of affected users. And than that number(s) on a tile.
Answer by Kris Z. ·
I'm not exactly sure whether you look to create a report for 1364 server hosts, each of them having its own clients, or for 1364 client hosts, each of which connects to an undefined number of servers, or something in between, and whether you need one number summarizing it all or one number per each of 1364 entities. Each of these perspectives makes of course sense for specific use case.
In any case, the way to approach this requirement is to use metrics that represent:
Affected users (availability)
Unique client IP addresses or Unique servers
Depending on the requirement, you may or many not need to adjust the way how users are tracked. NAM doesn't count number of affected client IPs, only number of affected users. When more than user sits behind a client IP address (e.g. behind NAT), number of users may be higher then number of client IP. And vice versa, when user's station changes IP address (e.g. as he moves from location to location), we may see one user, but two client IPs. When NAM doesn't see user name, it may use client IP as a user name or leave it blank. Unique and affected user counters depend on these settings.
Let's take an example where users are the focus. In this situation, I have number of user affected by availability issues counted and I have Failures TCP counted. So I can filter out only those servers that return TCP failures and then see how many users were affected on those servers. Voila - number of users affected by TCP issues is presented as a single counter.
I can aggregate it along any dimension, like software service type, analysis type, location, or provide a single global number. In any case, unique users counter counts unique users, so if e.g. user can't connect anywhere, it will still be one affected user.
You will need to experiment yourself with the best way to represent it for your client.
Hope this helps