Automation is as old as mankind. Since thousands of years we are working on a better life with less risk for our species. We invented simple tools and complex machines, we overcome physical distance on earth and created a massive network with global reach for data exchange. All that happened somehow in the past… But, nowadays people complain about automation. That is a narrative actually existing specially in the IT industry (funny thing is that we try to automate business processes with “digitization” and point our fingers on those workers afraid to loose there jobs).

Automation will soon cover most of the basic and easy processes in IT. That’s true. It will probably even take care of our service landscapes down to hardware in our data centers. Is that a bad thing? Think about your IT job, did you like(ed) to increase file-systems on request of your application teams? Do you want to go back to good old times when we installed patch bundles on our boxes during the weekends maintenance windows? In fact automation will cover such easy tasks and let us focus on more intensive and creative work. That can lead to more interesting jobs. Honestly i can not think of an option to get young potentials into your team when the job you can offer is mainly about standard administrative tasks.

Another aspect why automation is our friend is scaling. Did you ever installed 50 OS instances manually? Yes? It’s doable but not funny or inspiring at all. Did you tried 100? 500? Or even thousands? It is not doable cause as soon as you finished the last, you can start upgrading the first again. With micro-service architectures and cloud resources in mind that scenario becomes even more significant. Modern application architectures can most probably not get implemented, deployed and maintained without automation.

Other than this major reasons, we potentially get further values from proper automation like:

  • Speed
  • Quality
  • Predictable results
  • Standardization
  • Security
  • Fun
  • …

Is “Automation the new black” and in all cases the answer? No, please not! There are reasons not to start with automate things like the well known term “shit in, shit out”. Bad processes or architectures can not be fixed with automation. It might be smarter to first start a activity to fix them before you automate aspects. Popper planing has to take place to manage expectations. Always start small with potential low hanging fruits, automation is not a easy task, it will bind resources and take time to be done. Stakeholders might not accept that. Better be prepared with some fast results. The power of automation is understood, but as always huge power comes with huge responsibility. A single human error can be replicated unbelievable fast and end in manual correction of thousands of instances. People, documentation and lean design shout be part of your core believes in any automation project. You should also pay attention to your tooling. Instead of proprietary monolithic monster applications that demand so called automation engineers that do not understand the problem they automate can become a big problem. Not talking about find additional experts to scale such implementations. Do not. ;-)

And for those not reading one of my posts for the first time. Don’t miss your people/your team. Without a buy-in of your team you will fail.

Hope my thoughts are not to negative and help you think about further potential in automate aspects of your business. Do you have other experiences? Did i missed an aspect? Just let me know…