Showing results for 
Show  only  | Search instead for 
Did you mean: 

This product reached the end of support date on March 31, 2021.

CAS Scalability Suggestions for Intranet Apps having total of 180K Users

Dynatrace Organizer
Dynatrace Organizer

My customer's env is as follows and I seek guidance on which config/settings should I make in CAS (v12.3).

About the customer env:

  1. Dc RUM is monitoring a Data center where Apps are accessed only by Employees (for Corp IT). There are two Data centers and one AMD per DC.
  2. All Apps are accessed over http/https
  3. Total No of Apps = 20+. Except for 1-2 Apps which are accessible from Internet (using VPN), the rest are accessible in intranet only.
  4. Total no of employees = ~180K (180000). Emps login with their network userid/pw.
  5. Employees are located in one or more office locations world-wide. A few employees may access the Apps from outside office using VPN.
  6. All Employees do not access All Apps everyday. A couple of Apps are heavily used and are accessed by 20-40K employees everyday.
  7. All employees access Apps only after logging in. OpenAM based SSO is enabled for all Apps.

About Dc RUM deployment:

The CAS is running on a 64 gb Win 2012 Server VM and has SQLServer DB on same VM. Assume is sufficient disk space, vCPU's etc.

There is only one CAS (stand-alone).

The AMD is running on recommended physical hardware (HP Server) with 64 gb ram, sufficient CPU and disk space. There are two AMD's (because there are two Data Centers)

User names will be be extracted (from HTTP payload) for all Apps and reported in CAS.

CAS is configured to use the disk option to store user data with default values for disk space.

My Query: For Dc RUM to work effectively in this env, what scalability options should we enable in CAS? I have enabled the below settings and request someone to comment and suggest suitably.

Track Users with Identifiers, aggregate others(ISP mode)
(See attached screenshot)


Dynatrace Pro
Dynatrace Pro

For the number of users seem right. You can handle this. The only problem you can have is if there are too many diferents operations on the application that you are going to monitor. You can protect for a problem on this (if you have too much) limiting the URL autolearning on the softare service configuration and on the server limints on the capacity settings on the CAS

by the way if you are going to monitor all the IP, the check on the count on Ip that are going to be aggregated is useless, you are going to take all. I'll try to avoid to put the, I prefer to put the IP ranges that the your customer enterprise has and left the count IP for the aggregated user, just in case in the future your customer decide to inject the traffic for internet user