There was a time when the need for monitoring applications was really limited to the system administrators and application owners. Makes since when you think about it - they're the ones that caught all the heat when the applications went down so they had a vested interest in knowing their status. Application monitoring was simple back then. You just monitored the device that hosted the application and the process or services on that box that was the application and you were pretty much good to go. While it's still true that system administrators need a way to monitor their applications - they're not the only ones anymore and applications today are very different...
When we think about applications today, what we usually mean is the user's perspective of the application. For instance, I have users that leverage SalesForce.com. To them, "SalesForce" is an application. They launch the application's UI, in this case a web browser, and expect the application to appear and respond to their commands. To me, SalesForce is a service that involves applications, servers, databases, SANs, routers, switches, WAN accelerators, firewalls, bandwidth, and a whole lot more. These components have all got to be working together in order for the application to be available to my users.
Additionally, because the application relies on these various components, the people that need access to this information is much broader than just the application owners and system administrators. Network engineers, DBAs, infrastructure specialists, security professionals and just about everyone that supports the IT infrastructure could be involved in solving a problem with this application.
This makes it easy to see why device-oriented monitoring isn't a great fit for today's applications and services. When I monitor an application today I want to be able to see all of the different components that could affect that application on one screen and rolled up into service availability status for the application as whole. I want to see NetFlow data so that I can monitor congestion and see individual users' application traffic. I want to see how busy the routers and switches are that support the application. I want to see relevant syslog messages combined with Windows Server events. I also want to see status and perforance information on the aplication servers, SAN devices, and database servers.
This week here at SolarWinds we released the Orion Application Performance Monitor (APM) version 3.0. APM 3.0 does some really cool things. First off, it allows you to build application views that meet the requirements that I outlined above. Secondly, it integrates with our community site Thwack.com so that it's easier to share custom application monitoring templates that have been contributed or that you've written yourself.
You can of course download a free copy of Orion APM to try out for 30 days. I've been using it here in the office since before it went into beta and I've been really happy with it thus far.
Follow me on Twitter