University IT departments are charged with a lot of responsibility. They have to ensure a wide variety of services for faculty and students, they truly need to maintain 24/7 availability for critical services and they often play the role of a service provider as well. At Seattle Pacific University, Brendan McGree works as a system engineer responsible for monitoring servers, services and applications hosted both on campus and in the cloud.
McGree was tasked with managing the department's network monitoring software. His department had been using both Microsoft SCOM and PRTG Network Monitor, but when he took over, he realized changes needed to be made. SCOM had created a variety of problems and extra work, and the PRTG instance in use was a bit out of date. Ultimately, he decided to upgrade PRTG and became the in-house guru on how to use the software.
"One of the biggest things I liked about [PRTG] personally was that it doesn't have to have agents installed, whereas SCOM, especially if you're running anything not on Windows, you need an agent," McGree said. "PRTG didn't have that requirement on anything, which is awesome."
Getting monitoring into shape
Once he upgraded to the most recent version of PRTG, McGree quickly became the in-house expert on monitoring. He took advantage of PRTG's agentless structure and wide variety of sensor types to monitor the universities rapidly growing number of virtual machines. At its peak, the university had roughly 200 VMs running, with PRTG monitoring them all, along with physical servers, webpages, network traffic, CPU and RAM availability, and extensive local computing statistics, many delivered in XML format through custom sensors and Powershell scripting.
McGree is able to use custom sensors to generate critical data on the Nimble storage array networks (SANs) the university uses. Using the template and some scripting, McGree is able to generate reports on each data store and place them in their own separate channel, with alerts set that will notify them whenever one data store has 500 GBs or less free space remaining. Through creative use of custom templates and channels, McGree is able to monitor the entire SAN with one sensor and 50 channels, as opposed to the old days where it would require an agent and monitoring of each data store.
The university benefits most from PRTG in terms of notifications and understanding problems as they are occurring, before users feel them. McGree visited another university IT department to learn more about their Nimble implementation, and was surprised to find out they didn't use monitoring at all.
"The big thing that stood out at least to me is that they don't use a monitoring service, they just wait until students or professors send in complaints to the help desk," he said.
With PRTG, McGree and his department are often notified of problems before the bubble up to the user level, and have seen a decrease in the number of complaints sent to the help desk. And, even if the team is not in the office, they track complaints through email notifications and the PRTG iOS app.
"What PRTG has allowed us to do, especially if things go down in the middle of the day, is we can get on it and be working on it in minutes, and a lot of times even if it's a super popular system, we don't even get a complaint," he said. "It's nice from a customer service perspective, and it definitely saves us time."
Additionally, McGree said that the university rarely needs to use Paessler support, but has had nothing but good experiences with the team. In fact, he's become such an expert on PRTG that he serves as an internal support for the rest of his team. "If something goes wrong with PRTG they come and ask me, and it's great that nothing ever really does," he said.
"What PRTG has allowed us to do, especially if things go down in the middle of the day, we can get on it and be working on it in minutes, and a lot of times even if it's a super popular system, we don't even get a complaint. "
Brendan McGree, System Engineer, Seattle Pacific University