Customer perception for systems admins

When I was first learning system administration, a senior admin told me “Good administrators are invisible”. I did not really understand that at first, but now it’s one of my prime tenets. For the most part, we become visible when things break.  If you have to be visible, manage your visibility.

The first part of being invisible is planning. Always have a backout plan, and always plan your disruptions around times of low usage. These things require a mixture of common sense and data analysis. Monday mornings, plan on more email to flow. Likewise Friday afternoons. Unless, of course, Friday is, say, Christmas, in which case plan for the rush on Thursday. Data collection and parsing is your friend here. My personal favorite is Munin, which I plan to write a better article about soon.

The second part is anticipation. Anticipation is largely tied in to planning, but is much more organic. You have to have a holistic, insightful view of your customers and know what they will want before they do. This is obviously not easy, but again common sense and experience are critical. This is carries one of the exceptions to being invisible: Get involved with the developers and testers. Ask to be included in high-level planning sessions. You will undoubtedly ask questions and raise points from your POV that they simply might not think of. Don’t make the mistake of just bringing problems- try to have solutions as well. I don’t know how many times I have seen projects hamstrung because nobody thought about backups, or scalability, or other things they could have caught with a broader planning team. The bottom line is that making yourself more visible here can make you less visible later.

The third aspect of invisibility is response time. Response time needs to be as fast as possible, preferably fixing problems before anyone reports them to you. Monitoring is definitely your friend here. I prefer Nagios, but you may wish to look into Zenoss as well. Monitoring should be thoughtfully laid out and should notify you, not inundate you. Noisy monitoring systems quickly fade to background noise- make sure it only alerts you when something is seriously wrong, and then sparingly. Notify key customers of the outage, and kep them updated when you reach milestones. If you can designate a person to do this, do so. That way the admins can focus on the outage and the designated contact person can translate progress and manage the customer experience.

The fourth part is managing your visibility. We’ve touched on this above, with outage notification, planning, and becoming involved in the development process. The SA team should always come across as professional and approachable. It’s a hard balance to strike, for sure, but it will help you in the long run. Be sure that your communications, whether verbal or written, are informative and well thought out. A few bad emails can taint your perception, and managing that perception is a very real part of your job. Admins need to be trusted by their users- trusted with their sensitive data, and trusted to keep the systems running. Don’t let a perception problem cause you problems in the future.

If you can manage to accomplish these tasks, you’ll be well on your way to putting a friendly and reliable face on your team.

Leave a Reply