ServersCheck is a powerful monitoring product that allows IT administrators to keep tabs on server and overall infrastructure health.
For a long time, I've been looking for a product that wasn't horribly expensive that provided very simple at-a-glance operational status complete with graphs, gauges, and pretty colors. I also wanted a product that provides the capability to drill deeper into the details when necessary. With ServersCheck 7.0, I think I found it.
Installation and configuration
You can download a free, 15-day trial of ServersCheck from TechRepublic's Software Library. Once it's downloaded, you can click on the Start Configuration wizard option on the Welcome Screen (Figure A) to get started.
|ServersCheck welcome screen|
Use the configuration wizard to perform the initial setup of ServersCheck in your organization. By default, the ServersCheck services run in the context of the localsystem account. While this scenario is great for services that stay local — as in, the service does not access other computers — as soon as that service tries to get to something on another computer, the service needs credentials that can span systems.
On the screen shown in Figure B, you should provide the username and password for an account that has administrative rights over every system you plan to monitor.
In many organizations — particularly large ones — no one person is singly responsible for every infrastructure asset that requires monitoring. For example, in a lot of companies, the guy who administers the servers is not the only person who will be contacted should the data center overheat.
By using teams, (Figure C) you can have ServersCheck notify multiple people in the event of a problem. By creating different teams for different problems, you can make sure the right people are notified, which helps keep your SLAs working.
Although you can individually add each of your servers to ServersCheck, doing so can be a hassle. In most companies with well-defined IP addressing policies, servers are assigned to a particular range. During the ServersCheck configuration, you can define a starting and ending IP address, and ServersCheck will look for any devices that might exist in that range. Further, ServersCheck provides you with the opportunity to automatically create a "ping rule" for any devices it locates. A ping rule is simply a rule that periodically pings a monitored device to make sure it is still alive. If the device fails the ping test, ServersCheck can alert a particular team.
ServersCheck can use either your organization's SMTP server or it can use its own. The information you provide on the configuration page shown in Figure D determines whether ServersCheck can use its own SMTP server. I recommend that you do everything you can to use ServersCheck's internal SMTP server. After all, what do you do if your organization's SMTP server fails a check? How would administrators be notified?
|Designate an SMTP server|
ServersCheck doesn't have to be administered from the server on which you installed the software. You can use any Web browser to access the ServersCheck server and make configuration changes. As you would probably expect, ServersCheck's configuration is protected by a username and password. You can add other users after the configuration wizard is finished.
After you complete the configuration wizard, go to the main screen, where you will see ServersCheck's initial dashboard view (Figure E). In this screenshot, you see a single monitored server — the local system. Had ServersCheck located other servers during the configuration wizard, those servers would also be listed on this screen.
Note that Figure E provides a wealth of information. The big green circle indicates that all is well with the monitored devices, and ServersCheck even calculates your "current service level." Further, ServersCheck reports the number of up and down conditions, as well as which conditions are "iffy" (orange). In the lower half of the window, note the gauges that are associated with the monitored device. These gauges give you an at-a-glance way to determine the device's health.
Once you have a reasonable amount of information about your devices, you can begin to perform trend analysis. A trend analysis can be very useful in helping to plan future disk space needs, to identify possible upcoming CPU performance problems, and much more. To create a trend report, choose the monitored item on which you would like to report, name the report, and choose the dates and times that should be included in the analysis.
The graph in Figure F shows the CPU activity of my GoodLink server. This server is not heavily utilized, but this graph could point out anomalies that need investigation.
From the main screen, if you click on a device, ServersCheck provides you with a plethora of information about that device. In the screen shown in Figure G, you can see that the local system CPU is running at 1 percent with 75 percent of its memory available. You can also see this information trended over time in the graphs at the bottom of the window.
For each device you monitor, you can add a huge variety of rules to determine the asset's health. You can perform a typical ping check, run a tracert against a device to determine the path that is being used to access the device, and also monitor specific aspects of the device. For example, if you're running a Web server, a ping check tells you that the server is up, and most reasonable monitoring programs will allow you to watch the Web server service to make sure it stays alive. But how do you know if your page content is actually working?
Simple! Create a dummy page on your Web site just for monitoring and then create a ServersCheck rule that looks for specific text to be returned. If the text is returned, your Web server is serving content.
How often should the new check be performed? For some checks, you want a status update very quickly and for others, a longer time between checks is appropriate. Be mindful that you don't overload your systems or network with checks that run too often.
For my new check, I opted to watch a specific server service to ensure it doesn't use too much CPU time. In this example, shown in Figure H, the service is running on the local server, so credentials are not as critical as they would be otherwise. Bear in mind that for remote checks, you must provide credentials for a user account that has rights to gather the information requested.
|Setting up a monitor rule|
On the wizard's final screen, you can specify the conditions that warrant an alert and how the alert will be made. Note that ServersCheck provides a wide array of alert options. I recommend that you don't rely solely on e-mail. After all, if your Internet connection goes down, how would the software alert you?
As time goes on, you'll probably add more servers to your environment and will need to add them to ServersCheck. Provide the device name, IP address, description, OS version, device type, and the username and password for an account that can scan the new device.
Take note of the Minimum Service Level entry (Figure I). This entry is used to gauge the seriousness of a problem on the device. For example, if you're doing 10 checks, and one fails, your service level is at 90 percent. If this service has only an 80 percent service level, you're not considered "down."
|Minimum Service Level|
With the credentials you provided, the new device is scanned and appropriate monitoring options are returned. In my example, I've added a Windows domain controller, so Windows Health Checks are available, which include CPU monitoring, memory monitoring, and disk space monitoring. You can also perform TCP checks to make sure that specific TCP ports remain accessible on the new device. Finally, you can watch specific services to ensure that they stay available as well.
If you have a system with a problem, the dashboard will show that problem. In Figure J, the "details" option for the local system indicates "Error: Query returned zero processes." In this case, I was scanning for a service that did not exist on the server, but this does show you how easy ServersCheck makes it to figure out what's wrong.
As you add checks to a device, they appear on that device's "device view." If a server is down, you have a problem, and ServersCheck goes red to let you know about it. Note that the server named AIS is currently down (Figure K), a fact indicated front and center. The pie chart has also changed to reflect this device's down status.
|A server is down|
The bottom line
In my testing, I found ServersCheck easy to use, and it definitely provides the visual cues I wanted. Using ServersCheck, we can put a large monitor in the middle of our IT suite and track our assets' status. I like ServersCheck enough that I plan to adopt the product in my organization as our primary monitoring tool.
However, I did run into a few problems. First off, I was unable to monitor one of my servers due to a problem with Windows Management Instrumentation (WMI). I'm sure that's not related to ServersCheck, but so far, ServersCheck help has not been helpful. To place a phone call to the company, you need to fork over $75. This is not something I'm willing to do when I'm evaluating a product.
I also tried to post my question in the forums, but you need a paid license to do that. When I tried to register for the forums, I received a registration response that said "Your registration contained an error." I forwarded the message to ServersCheck's support e-mail, but the message bounced. At ServersCheck's price of $299.00, I'm definitely willing to risk it, and once I pay up, I'll be able to get the help I need.
You can download a free, 15-day trial of ServersCheck from TechRepublic's Software Library.