DX Unified Infrastructure Management

I just want need opinion regarding the UIM implementation, we have plan to setup our production UIM servers with 3 primary hubs running in a High Availability, 5 servers on secondary hubs and mssql server in a cluster environment. This UIM will monitor almost 1600 servers.

In the CA docops documentation have advise to use Mssql Server Enterprise Edition for the database system because it is supports table and index partitioning.  Please refer to the link “https://docops.ca.com/ca-unified-infrastructure-management/8-4/en/installing-ca-uim/pre-installation-planning/install-and-configure-your-database-software#InstallandConfigureYourDatabaseSoftware-MicrosoftSQLServer”

In this case, based on the size of the implementation, do I need the table and index partition to support for this implementation? Can I use Mssql Server Standard Edition?

It is normally recommend that partitioning should be enabled where possible even on a smaller database, it will improve the performance of database for sure and makes it easy for the maintenance. Looking at the size of your environment  (1600 Servers) and the number of metrics you would be collecting for each server and depending on the data retention settings the database size might grow to a good extend .

Rather than looking at the number of servers you would be monitoring , look at the number of metrics which you would plan to collect from each server and how long you plan to retain the data  which actually defines the size of the DB growth .

Try the calculations based on the number of metrics planned to be collected , it might be a good idea to go with Enterprise edition 

"Try the calculations based on the number of metrics...", is there a sizing calculator or table to estimate the database sizing?

Not sure there is something available for customers to use (few things I know by past experience) , take a look at the below and it might be helpful. An environment of 1600 servers might fall under large company

Prepare Your Server Hardware - CA Unified Infrastructure Management - 8.4 - CA Technologies Documentation 

This sounds like way overkill for just 1600 servers. We have double that and are running standard SQL and have our DB running on Extreme IO. We have a schedule to re-index the db every two weeks.

Our Primary Hub is running Windows and we have 3 Tunnel servers running on CentOS with 100+ client hubs reporting up into them.

The main part here is to adjust your RAW data retention. The default of 180 days is just way too much. 6 months of RAW data collection is not needed even in our MSP environment. Our SQL DB shrunk from 1.5TB with 180 days of RAW down to 240GB with 14 days of raw, 140 hourly, 490 daily roll up.

If you keep the db small as such, you can run the scheduler to re-index the DB as needed. If you have the SQL enterprise license then use it and it can re-index as needed with the data_engine option checked off.

Hi All,

I have a question, Would it be possible to see the daily growth in UIM database for monitored servers ?. if yes, could you please share me the table name and sql query.