IT administration is changing—expanding, to be precise. We can no longer think of just “our part” of IT, “our part” of the network, or “our part” of the help desk. In order for agencies to efficiently and effectively manage IT, administrators must have visibility across the entire application stack, or “AppStack.” Only with visibility to all infrastructure elements that are supporting an application deployed in production can they understand how those elements relate to each other.
This, however, is easier said than done. There’s a good chance there are things in your AppStack that you’re not even aware of—things that, for better or worse, you must uncover and track in order to gain critical cross-stack visibility. Meeting the challenge will likely require broadening traditional definitions of an AppStack to include things we traditionally thought of as out of scope.
What kinds of things? I’m glad you asked.
Not too long ago, application management was simple. well, less-complex at least. Applications were more limited in extent and sat on a mainframe, micros, or PCs with local storage. Today, the complexity of applications—with shared storage and server virtualization—begs a restatement of the meaning of an application to include the entire AppStack. Today, the AppStack includes everything from the user’s experience to the spindles and everything in between. Yes, everything in between.
As many agencies are moving to a hybrid-cloud environment, let’s take the example of a hybrid-cloud hosted virtual classroom system. Are HTTP services for presentation and data tier included? Yes. As are storage and virtualization, as well as core, distribution, WAN, and firewall. What about VPN and firewall to the Virtual Private Cloud (VPC)? Yes— Cloud has become an important component of many application stacks.
There are even things you might have assumed aren’t components of a traditional application—therefore are not part of the AppStack—that are, in fact, critical to uptime. Some examples are security certificate expiration for the web portal, the integration of durable subscription endpoint services, and the Exchange™ servers that transmitted emails and relayed questions. Yes, these are part of the AppStack. Yes, you should be monitoring all of these elements.
You will achieve the greatest success in application monitoring and management by expanding the definition of AppStack to include all elements of your environment. Encourage all the Federal IT pros in your team to broaden their perspective. Use your network monitoring systems to discover and document every element of the connectivity chain, identify and document the links between each. You will likely discover that there are more than a few previously unmonitored services in your AppStack you need start keeping more of an eye on. And in the long run, proactive observation has a tendency to decrease unplanned outages and even make your weekends a bit better.
