What is most valuable?
The features valuable for me are scalability, redundancy, and the wide range of probes available for just about any platform.
Another major advantage is the easy configuration management. When you define standard “base” monitoring templates and on top of those, define “differential” templates, having a tool that allows you to manage these hundreds (and even thousands) of templates in an organized manner is an absolute necessity.
CA UIM not only allows you to manage the templates, but the new MCS module allows you to dynamically assign them to groups. This means that any node/probe belonging to that group will automatically receive the templates.
Note: At this time, MCS does not yet support differential templates, but it is on the roadmap.
When this customer did the PoC, the competition clearly failed in this area. Furthermore, the event management part (alarm server and alert console) is feature-rich. It allows for some advanced, alert processing and correlation.
How has it helped my organization?
Nodes are now monitored in a standardized way, thanks to configuration management. We are a lot more pro-active to solve potential issues, thanks to event management.
Different teams are using different dashboards according to their requirements. For example, the end-user service desk uses a high-level dashboard with the state of the most important applications, printer malfunctions, etc.
A lot of the functionality is available out-of-the-box without having to script it, although scripting it is still needed from time to time. This means new objects/metrics are effectively getting monitored quickly without a lengthy development period.
What needs improvement?
I would like to see the retirement of the heavy client (infrastructure manager) in favor of the web-based admin console. It is close, but it is not there yet.
Support for the PostgreSQL database platform would be nice. At this time, you can only choose between Oracle and MySQL when running CA UIM on Linux.
As a DBA, I prefer PostgreSQL over MySQL. (This is my personal preference. By no means do I find MySQL a bad product.)
For how long have I used the solution?
For this customer, it’s their first implementation. However, I have been using CA UIM since 2014 and its predecessor, Unicenter NSM, since 2004.
What do I think about the stability of the solution?
We had no stability issues. The built-in redundancy allows for maintenance windows for patching.
What do I think about the scalability of the solution?
Scalability is one of its strong points. A manager server (hub) can manage a lot of nodes. Adding another hub is really straightforward. You just need to make sure you have plenty of storage for both the database and the primary hub.
How is customer service and technical support?
The technical support is very good. You still need access to the old support.nimsoft.com site for downloading new versions of probes. Overall, it is very good.
CA support did have a bad reputation 10-15 years ago, but they made a lot of effort to improve it.
Which solutions did we use previously?
The customer was using Nagios, and still is for the network part.
The main issues were:
- The lengthy development periods (because you need to script everything),
- The difficulty of having any custom reporting/dashboarding capabilities
- The total absence of decent event management functionality: This is something most cheaper monitoring tools anyway don't have
How was the initial setup?
The actual installation was very quick and straightforward. Of course, you can spend ages configuring/tweaking as the product has endless options. It all depends what you have defined and how complex your environment is.
Most of the time was spent defining the architecture: What probes go where, redundancy, tunnels for the DMZ networks, and capacity planning. I recommend spending as much time as required on this, as you will benefit from it later.
What's my experience with pricing, setup cost, and licensing?
It is definitely not the cheapest on the market, but you can save a lot of money by carefully making a list of which probes you actually need and how many of them are required.
Often probes can monitor multiple instances or are included in other packs, so you don’t have to purchase them separately. Make sure to be precise. Your CA representative can assist you with that.
Which other solutions did I evaluate?
The customer did a PoC with two contenders: CA UIM and Microsoft SCOM.
What other advice do I have?
Try to look beyond the price tag. You really do get a lot in return, especially when you have a highly heterogeneous environment.
Disclosure: My company has a business relationship with this vendor other than being a customer: We are a service partner for CA, Microsoft, and IBM.
Apr 30 2017