Last week I interviewed Joe Kline of Maritz. Joe is a Senior Infrastructure Specialist and manages the Network Management Group which is responsible for deploying new IT solutions, tools, upgrades, and more. Joe and his team use Web Performance Monitor from SolarWinds to manage hundreds of web applications and websites.
Jennifer: What SolarWinds products do you use to monitor your environment and how do you use them?
Joe: We have Server & Application Monitor (SAM), Web Performance Monitor, Network Performance Monitor, Network Configuration Manager, and VoIP & Network Quality Manager.
We previously used HP SiteScope and their Business Availability Center (BAC) products. These products were pretty expensive and the SolarWinds Orion toolset offered more flexible licensing options.
Orion offers a lot of flexibility for monitoring. I do like how flexible SAM is, we can pretty much do what we want to do. It is just basically limited by whatever we want to put our heads and minds together to work on. HP kind of limited you on what you could do.
We use Network Performance Monitor (NPM) for the core CPU, memory, and disk space for pretty much everything. We use Server & Application Monitor (SAM) to cover a lot of application specific monitoring like Windows services, processes, and some performance monitoring counters. We use Web Performance Monitor (WPM) to monitor all our web applications, which is quite a few.
Jennifer: Could you explain how you monitor your websites, what you’re looking for, issues you uncover?
Joe: We are currently monitoring about 315 websites with WPM today. That could easily double if WPM improves its scalability. And for SAM, we probably have about 25,000 component monitors deployed in roughly around 1,700 servers. We are several different business units operating under one big parent company so we have different application development groups supporting each of those lines of businesses. We deal with a lot of different architectures when it comes to web applications. We are very heavy on the IIS and .NET side but we also have a pretty sizeable installation of JBoss. Our business units under Martiz all develop applications that you probably use today if you have a credit card with any kind of points reward program. That’s the kind of things we do. We host a lot of these applications at our HQ in Missouri, but we also have onsite deployments and we are looking at venturing into the cloud.
We have hundreds and hundreds of those applications and SLAs we have to adhere to. Some of which, especially in the financial sector, are very strict with significant financial penalties if we don’t adhere to SLAs.
With our website monitoring, we are primarily looking at availability. We do look at performance on a limited basis and if we see a performance issue, we bring it to the business unit’s attention. For example, we look at every step in the transaction and try to put content matches in where it makes sense. We find issues with the websites every day, at least 10 to 15 alerts every day.
Jennifer: If you didn’t have web application monitoring in your environment, would you be getting a lot more calls?
Joe: Over the last 4 years, we’ve really matured as a Network Operations Center in how we monitor everything. I think our customers that use our services that we’re monitoring applications for comfortably rely on us a lot more than they used to. Monitoring web applications 5 years ago was kind of an afterthought. We did it as requested and now we pretty much monitor everything that we know about.
The obvious benefits are if we’re notifying them (clients) of issues before the client does. We provide the business with a lot of reporting (internally built) driven off the Orion data, largely availability and performance on a monthly level. We also take our change management infrastructure changes, and correlate that with downtime events and manipulate the availability metrics based off those windows, so we can have more realistic SLAs which exclude maintenance windows.