alex_aldo - Fotolia
In 2014, Adrian Dupre knew his network was going to the dogs.
Today, it still is.
But that's OK. Dupre, vice president of IT services at Seattle-based Trupanion, the nation's largest provider of pet health insurance, has carefully laid a framework to ensure that his network will always go to the dogs -- and cats -- covered under the company's plans.
"We needed something in-house and something that would let us pick up anomalies" as they occur, Dupre said, about Trupanion's deployment of Paessler's PRTG network monitoring system software.
Fueling the implementation of in-house software was the decision to transition Trupanion away from an infrastructure-as-a-service (IaaS) model to a self-managed network, Dupre said. The old network monitoring software -- LogicMonitor -- was no longer sufficient, and if Trupanion lost its Internet access, it would have no way to evaluate network performance of servers based in data centers at Santa Clara, Calif., and the Seattle area.
"We needed something more industrial-strength and something that wouldn't break the bank," Dupre said.
Ease of management needed to streamline operations
Ease of use, meantime, was another important consideration. "With IaaS, we didn't have a lot of system administrative staff, and [despite] 100 systems and half a dozen business-critical services, we are still operating very lean, so we wanted it to be easy to manage," Dupre said. "Literally hours from the time we installed the software it started uncovering things we didn't know about. Just the basic out-of-the-box monitoring was great, but more important [PRTG] bought us time as we transition our management strategy." In addition, PRTG permitted Trupanion to better understand the compute and server resources required to support new products and applications.
Having that foundation, Dupre said, will enable his IT staff to more precisely monitor all of Trupanion's business services. "We can now say, 'We have 4% [capacity] left on this server and we can set up thresholds and customize each of our servers so that [its performance] is tailored to our business.'"
Trupanion apps support payment, enrollment functions
Among the most critical IT functions is support of trupanion.com, through which Trupanion captures half of its enrollments. "Our customer-facing website is absolutely critical," Dupre said.
The company also maintains a suite of applications, both custom and vendor-supplied, that enable it to track the more than 200,000 pets it covers. "We have billing, enrollment and monitoring of policies, but we also have intake of claims and payments, so our internal management suite is like a [manufacturing enterprise resource planning] system," he said.
Adrian Duprevice president of IT services, Trupanion
Another custom app, Trupanion Express, handles payments and approvals. The software is engineered to integrate with veterinarians' patient information systems. Trupanion promises to turn around claims processed through the app within five minutes, so performance is critical.
Trupanion is just at the beginning of deploying PRTG. "Right now it's basic monitoring on our farm of servers" -- spanning domain controllers, Exchange, SharePoint and SQL among other apps -- "and we're still at one probe right now, but we have significant ambitions," Dupre said.
Trupanion is also using PRTG as a diagnostic tool, Dupre said -- for such myriad functions as pinpointing possible phone problems experienced by teleworkers or for meeting Sarbanes-Oxley financial regulations.
"All of our business processes have to be under control, that your backups are working properly, and a lot of these are running on Windows hosts and writing to Windows events logs. Rather than writing additional code and designate a person as a log jockey, we have PRTG look for certain messages that represent failures. We can see what's logged and we will essentially have PRTG log a help desk ticket any time we have an unusual event or a failure in our batch jobs. It really increases our efficiency," Dupre said.
Next up for Trupanion as it further taps PRTG's feature set: baselining network performance. "We want to look at each of our applications that we are running and be able to say, 'What is normal operation and what do we care about versus what we don't care about,'" Dupre said. "A lot will be tuning and the next part is how we use [PRTG] to automate some of our routine monitoring. Can we turn PRTG into a central operations center to show us what's going on? I think what we will be doing is looking through the sensors that are published and the new ones and then we can really see where we can use them to add value to the company and really tie our monitoring into business performance as opposed to beanie-head propeller-twirling performance."
Understanding app performance management
Preparing for data center management challenges
SDN monitoring challenges percolate