What is our primary use case?
My primary use case for SCOM is to monitor service availability and performance, such as operating systems. We also integrate some Linux based operating systems to monitor our databases. We also monitor the Microsoft Exchange. We are having some difficulties in the case of the monitoring a couple of our networking devices, so I wouldn't say that monitoring networking devices is also part of the primary use cases.
I also have Internet Information Server and Application Service from Microsoft Monitor.
What is most valuable?
Availability monitoring is the feature I have found most valuable, as well as the capacity and ability to send notifications. There is a mechanism to set up a notification from the SCOM and whenever there is a drop in the availability the notification alerts not only for availability but for other issues as well.
You can align thresholds according to the speed of your environment and you can have a threshold related notification, which is one of the useful features.
What needs improvement?
In terms of features that could be improved, I would say the agent integration into the operating system. We are having difficulties integrating Linux into some of the networking devices. We have not seen the collected data so it makes it challenging.
I would also say that agentless monitoring needs to be included. Something like this is pretty difficult if you don't have a particular agent.
It's not so easy if you have to use something like a proxy to implement a work around. They should include a solution for discovering devices and something like an agentless monitoring solution for a particular device - just to understand what your environment looks like.
I'm not saying that they should provide all the information for the device, but at least availability and partial monitoring based on SNMP. Because I know that other solutions have it. Maybe Service Center Operations Manager has already provided those things in the latest version, but I'm not familiar with it.
For how long have I used the solution?
I have been using Microcoft SCOM for roughly five years.
What do I think about the stability of the solution?
Our whole department that maintains SCOM is five people, so the IT department is pretty small and it's relatively integrated. Plus, we have only two dedicated System Administrators. Basically, a few people are doing many things.
We are doing the usual maintenance, patching, and updating. Alignment and configuration are also needed because we are trying to support a higher version of the application, Exchange 2016, with the previous 2012 version of Service Center Operations Manager. It's quite difficult and some type of maintenance is performed internally as we try to align as much as possible.
Generally speaking, SCOM is stable from an operations point of view. Once it's set up and established and all the configuration is in place, there is no significant amount of time needed for stability or to support the availability of the solution itself. It's relatively simple to support.
What do I think about the scalability of the solution?
In terms of scalability, we increased the amount of supported devices from 50 to 150, but it's not a significant scale. I couldn't say yet how it would behave for a thousand or 2,000 devices.
We have a pretty simple setup for SCOM, we are not using any kind of clustering or virtualization.
How are customer service and technical support?
Microsoft support can vary. It's great for people that are used to other types of support. Meaning, software vendors usually provide a little bit more focused and dedicated Microsoft support teams. For general support, it's okay from my perspective. But to really understand the deep and the intense nature of your issue, it's a little bit tricky to get to that level.
Overall, technical support needs some alignment. The technical support should be separated from general support. They should start with the general support for people that are end-users or a single user in a simple environment. Then, to reach to the real technical integration-related support later because it's a little bit tricky. Usually, we use our integration partner as someone that is really dealing with those kinds of discussions with the vendor. But if you go directly as a single enterprise, it will be a little bit challenging.
Which solution did I use previously and why did I switch?
In my previous work I had experience with Network Node Manager, a new solution from HP. It's more focused on agentless monitoring. Also, there is a capability for a monitoring device without any kind of agent installed, which is a little bit more professional. It's not the general solution for monitoring, so I wouldn't say that are comparable. My perception is that Microsoft is just focused on the software stack layer they provide. While Network Node Manager is more of an umbrella, it's much more focused on enterprise level devices and environments. So they should not be compared.
How was the initial setup?
As far as I can remember, the initial setup was pretty simple, but we started with pretty easy devices to monitor. We started with a couple of servers and almost all of them had the same version of the operating system. So it was simple to figure out how things were progressing.
I would say it took a couple of days or a week for the whole setup. It took about five days to install all the servers, and installation and all the integration, and the checks probably took a week.
What about the implementation team?
Initially we implemented on our own. But for a couple of integrations, we used integrators for support, especially for particular application servers and installations, Exchange and partially for SQL Server.
What other advice do I have?
My recommendation to anyone looking to use SCOM is that they should start with the latest version because it's pretty difficult to update later. From an operational perspective and regarding migrating many devices and whole environments, it's better to start with the latest version for Service Center Operations Manager.
On a scale of one to ten I would rate Microsoft SCOP a seven.
Which deployment model are you using for this solution?