This week we are attending the Microsoft Management Summit in Las Vegas where we will be showcasing our products and demonstrating how they complement Microsoft’s System Center Operations Manager. Over the last couple of months I’ve been gathering feedback from customers and administrators on their experience using SCOM for monitoring and reporting within their messaging and collaboration environments and I’ve prepared a summary of what I’ve found which I’m going to share with you today.
Alright, here are the four main things that everyone I talked to agreed were important for monitoring their messaging an collaboration environments.
- Simple - Low impact and easy to deploy, configure and manage
- Relevant - Alerts should be important, not overwhelming
- Performance oriented - Not only the servers, but user experience as well
- Reporting – The ability to report on server, environment and service level performance
So how did SCOM rate?
Deployment - For one, most of the individuals I spoke with indicated that deploying the management pack for Exchange was fairly straightforward and simple, that is though, if SCOM is already deployed in the company to begin with. If SCOM is new to the environment altogether it requires a comprehensive deployment plan with input from multiple teams for integration meaning that simply monitoring your messaging environment can be a little way out.
Configuration - From my discussions I determined that the configuration of SCOM for Exchange required the most amount of effort. When installing SCOM all components and features are installed automatically, even if unwanted or unneeded requiring a great deal of configuration to ensure accurate, relevant monitoring. There are a number of out of the box preset once installed but it is likely that these will not necessarily be suitable for your environment. According to the individuals I spoke with this can take a great deal of time to complete, delaying the expected benefits of the product. Once configured however, most people reported being fairly satisfied with the experience.
Ongoing management – A couple of things stood out here, the agent required on the monitored server, and the potential need for retuning. The SCOM agent really had two basic drawbacks that were mentioned.
- Monitoring and reporting should have minimal impact on the monitored systems. An agent on the server requires resources and can impact service, especially if the agent conflicts with the monitored server.
- Agents require occasional upgrading which is another component that needs to be tested and a plan needs to be put in place to actually install. Agentless tools do not require this and in this case can be easier to manage.
One other thing that was brought up was retuning. If you experience any changes in operations, such as new servers, merging companies or expansion your presets may become irrelevant requiring you to potentially go through a painful process of retuning and hundreds of nuisance alerts.
Administrators were especially vocal about the need for alerts to mean something. I touched on it above, SCOM does require a fair amount of tuning and the periods in-between can produce a large amount of nuisance alerts. What’s a nuisance alert? An alert that is generated prompting an action from the Network Operations Center (usually a call at 3 in the morning) that actually has no impact whatsoever. Administrators want to know that when they get an alert something is actually impacted, or even better, something is about to be impacted if you don’t take action.
Historically monitoring tools have focused on things like CPU percentage and memory utilization, I know that’s a fairly simply summation of what monitoring tools look at, however, the point is, that administrators want something more. They want to be alerted when there is a potential user impact. CPU and memory is important generally for capacity and planning, occasionally a high CPU alert (depending on set threshold) can bring awareness to potential user impact. What’s needed is a system that alerts you to a potential or imminent user experience issue. Having this information allows administrators to make quick decisions on how to maintain service levels. General opinion on SCOM is that it provides a great deal of statistical information but it leaves out the user experience component. There are some tests that SCOM runs that begin to address this but only from an availability standpoint.
The impression I have from messaging administrators using SCOM is that it’s fairly limited. Administrators are looking for the ability to quite simply automate reporting, trending and forecasting eliminating any manual work. Having a great monitoring tool allows administrators to react to their environment, having a great reporting tool will allow administrators to proactively manage their environment and minimize the change of impacting user experience.
Overall SCOM for Microsoft Exchange provides a robust monitoring tool that covers a wide range of server level statistics. It is seen as a comprehensive tool but it lacks some of the user or service level monitoring that one would expect. In many environments the benefits of SCOM can be magnified with tools that focus on service related alerts, user experience and quality reporting.
If interested, below I continue this article highlighting some of the complimentary features of GSX Monitor and Analyzer.
GSX Monitor and Analyzer
Ok, now that I’ve reviewed our findings on SCOM (and really monitoring in general) I’d like to share with you how you can use our tools to complement SCOM and fill in some of the voids that exist. GSX Monitor is a powerful application monitoring solution that enables administrators and IT Managers to proactively and efficiently manage and maintain their entire collaboration environment from a single User Interface. GSX Monitor works in concert with GSX Analyzer. GSX Analyzer is a robust executive dashboard and reporting tool that gives administrators and managers throughout the enterprise the information they need at the time they need it. It can help to understand key trends and performance metrics in your infrastructure enabling you to prioritize and act.
How can GSX complement SCOM?
GSX Monitor and Analyzer can be installed, configured and alerting (or hopefully not) in less than an hour. The tools do not require any agent on your servers and thresholds can be configured in a matter of minutes. In case of upgrade you only need to make changes on the GSX Monitor station, this is an in-place upgrade.
GSX Monitor only sends an alert when your users or service are impacted or about to be impacted. This means that you’re a step ahead of the issue and you don’t need to sift through hundreds of nuisance alerts to get to what’s important. This will ensure that when you get an alert you know that action needs to be taken. Once you’ve been alerted, you can correlate any SCOM notifications to help resolve the issue.
GSX Monitor and Analyzer are centered on ensuring performance from a user perspective. It is one thing for a server to be up and running, it is entirely another thing for the service to be performing as expected. GSX Monitor will alert you in case of user impact and GSX Analyzer will allow you to plan for capacity and measure your performance against any set SLA. All of this centers on what’s most important, how do your users experience the service you provide.
In GSX Monitor you have access to all current statistics and the ability to view graphs on recent performance. In GSX Analyzer you will be able to view historical statistics, trend your growth and even forecast what’s expected in the future. This will allow you to pinpoint any negative trends, plan for capacity and stay a step ahead of performance issues. What’s more you can also view a dynamic snapshot of SLA performance in Environmental Health. Environmental health lets you set an SLA for a given performance indicator, then select a number of KPIs that are critical to achieving your SLA. If performance isn’t meeting the SLA, you can quickly determine which server is impacting it and in what fashion. For example, if server availability is 99.9% and you see that your only at 98%, you can quickly scan down and find the offending server. This server only has 80% availability, and what’s more RAM utilization is constantly at 99%. To improve SLA performance the solution is simple, add more RAM or remove the server from your environment, you’re now meeting the SLA.
GSX Monitor and Analyzer can help you fill in some of the gaps and ensure that you are delivering the performance that your users expect and at the same time provide a comprehensive reporting and analysis tool. Together System Center Operations Manager, GSX Monitor and GSX Analyzer serve all of your monitoring, analysis and troubleshooting needs.
Thank you if you’ve opted to indulge me and learn how you can use our tools to complement your SCOM implementation. If you’d like to learn more please click on one of the links below.