Part IV – 7 Habits of Highly Effective Network and Systems Administrators
Continuing with our series, 7 Habits of Highly Effective Network and Systems Administrators, the next “habit” we’ll cover is “Template Usage”. When I think of the word “template,” I’m taken back to 1978 and watching Star Wars for the first time. All those Stormtroopers lined up, identically, waiting for Darth Vader in the hangar of the Death Star. Pretty cool. Without going too far down the Star Wars Canon rabbit hole, it’s fair to say the Galactic Empire had a template for how their rank-and-file troopers looked, sounded, and behaved. Not coincidentally, this technique should be mimicked to make your IT team more effective in their day-to-day activities.
Template Usage: Why Go Down This Path?
At the core of every successful and effective NMS strategy is a common configuration for all the devices under management. Variables like SNMP strings, Windows credentials, and SSH passwords need to be known by your monitoring system to extract information. However, in all but the most minuscule IT environments there are potentially thousands of devices of every shape and size to track. Templates allow you to quickly and reliably set those variables on a mass change basis. Certainty of a consistent configuration affords your team the ability to stay on top (or ahead) of problem conditions that would otherwise send an angry mob of pitchfork-wielding end-users to your office door.
In addition to core settings like credentials, templates allow you to enforce uniformity among similar classes of devices. A common use case here is different types of application servers. Every IIS server runs MS-Windows, but every MS-Windows system doesn’t necessarily run IIS. Therefore, a template associated with just your IIS servers is the appropriate solution. Another use case is if your network is globally dispersed. Network latency to all your sites is a metric you’ll want to track and alert on if conditions degrade. However, the laws of physics tell us that latency will always be higher for gear that’s further away. Therefore, it makes no sense to give all your devices the same latency threshold (unless you’re okay with tons of false positive alerts).
Consider the following scenario: An HQ datacenter in Denver, an R&D facility in Bangalore, and distribution centers in Dallas, Orlando, and London. The correct solution here would be to setup a template that you attach to all the equipment living in Bangalore and London where the network latency threshold is 10% to 15% higher. How does this make your network and systems team more effective? They’re not wasting their time chasing down non-critical alarms. Instead the NMS tool allowed them to create an “International” template, which can be applied globally to all those devices quickly, simply, and reliably.
Of course, “templates” is a generic word. While my prior examples have associated it with the configuration of devices under management, that isn’t always the case. The overarching goal of template usage is consistency, and that applies to the way outbound alerts from your NMS look, too. Pretend for a moment you want to integrate your NMS into an ITSM trouble ticketing system. If the only access into that system is via SMTP gateway, then any alerts that come off your NMS system have to be formatted in a very specific and predictable format. Further, suppose you want your NMS system to dictate trouble ticket priorities that get opened for your engineering team? A predictable alert format will be imperative. It should go without saying that tailoring alerts before they even get to your staff will boost their productivity.
What To Templatize
You might be wondering, “Ok, great. You’ve told me why, but what are some concrete steps I can take to get started on this magical “Template Usage” path you speak of?” Since network and systems management (and IT infrastructures for that matter) are rarely a “one-size-fits-all” proposition, consider the following guidelines. You’ll need to season to taste.
- Perform an engineering assessment of your environment. Are there obvious delineations of gear where settings would be different? Server templates vs. network gear templates are an obvious example. However, can you break it down further? Do your Windows and Unix/Linux teams play nice together? If not, maybe they need their own templates with particular contacts and thresholds.
- Perform a business assessment of your environment. Previously, I only addressed templates with regard to configuration settings and alert formatting. What about report templates? Do particular users at the business/functional level of your organization require regular and predictable reports to do their job effectively?
- Perform an application assessment of your environment. Applications are the soft and squishy area that falls between your IT people who likely maintain them and your business people who use them. What are your mission critical applications? In this instance I’m specifically thinking about templates for application integration (e.g. alert formatting for NMS <-> ITSM), but it applies to other areas, too. For example, does your ERP system require certain parameters to be monitored that are important to both the user and technical communities?
How Now Brown Cow?
We know why templates are a good idea. On a basic level, we also know what to templatize in the environment. So, the last step in the journey is, how do we get started? The best answer to this question is to dig into the capabilities of your NMS system and figure out how to create templates based on all the steps just discussed. Hopefully, your chosen NMS vendor has the flexibility to setup templates on a type, site, and category basis. If those capabilities don’t exist, remember the ultimate goal with templates is to widely and predictably configure your NMS tool with as few “clicks” as possible. That is how your staff will make the best use of their time. Therefore, you’ll want to seek out global configuration utilities within your NMS application to accomplish this task.
When all is said and done deploying templates, as part of your NMS strategy, is a no-brainer when it comes to making your network and systems administrators more effective at their jobs. Will they become so effective that they’ll eventually have the power and wherewithal to rule the galaxy with an iron fist? I suppose that depends upon how full-featured your NMS application is, but probably not.