The Logic behind Automation

Today we’re speaking to Vlad Moraru about the thought process behind any Automation Project he begins. When it comes to automation, the goal is generally to simplify processes with the added benefit of cutting time, cost, and margin for human error. But preparing and implementing automation systems still requires a lot of thought process, so how does an engineer review automation when consistently working on new projects for enterprise companies, meaning those projects will have a huge number of teams review them in their lifetime? Here are some of Vlad’s thoughts on the topic:

What’s your approach towards automation when you get a new project?

I always have a thing where I like everything to look uniform, I can’t work in a messy environment. A lot of projects have new project teams on each section, and every time you get to it there’s been on average 4 or 5 teams and it can look messy with so much rotation and no consistency. I like things to look a way that is logical to me in my head and will be easy for any team that follows on the project to be able to understand, the way I’ve been taught, even in the current project I’m working on you can see the pattern the team I’m in takes because we agree and work well together. Everything has uniformity and no discrepancies. If everything is unsightly and it was up to me how I use my time I’d happily spend a week cleaning before I even start. From doing that, the automation comes so much easier and things just click.

What I do is obviously highly dependent on what tech is being used, in my experience, it’s rare that you come across a project that doesn’t already contain automation. People always attempt it even if it doesn’t always come out successful. The important thing is to notice the areas which could benefit from being reworked, for example, if there is a new tool out there that can replace the current work I will try to slowly migrate everything to be up to date, and that’s worth doing even if it takes a month but works better and more efficiently in the long run. I’d gladly put in extra work for a month because it’ll improve it in the end rather than sticking to a vigorous process for say 6 months but nothing significantly improves. In the end automating just makes sense and makes your life easier down the end of the line. 

How far do you think automation should go?

I think things should be automated to the point that you don’t do anything, a one-click, out the box, works by itself and walk away level is the ideal. My biggest issue with that is you have to trust what you’ve done, especially working in a team you then also have to trust others. By yourself, you know all the code and know where the issues are, in a big team it’s also down to them to hold the same standards and communicate everything, in my current team of AL engineers, I have full trust that everything they do will be to a level that I’m happy with, it comes down to how much you trust yourself and those around you, that’s how much you can automate. 

What’s your favourite automation tool and why? 

In my experience so far, I’ve mainly worked with Ansible and ARM templates. At the moment  I’m really enjoying ansible. I can waste hours looking at it and playing about it, but there are definitely other tools out there that would be good to play with, I just haven’t had the chance to fully experience them all yet. 

At what point does automation become more effort than there is payback?

That becomes effort when the people you’re working with don’t automate everything, for automation to work everyone has to automate their process. if it’s just me, sitting in a corner and automating away, what I’m doing doesn’t make sense. You are slowing yourself down and your team down because you don’t understand each other. On an individual level, I don’t think it’ll ever be more of a problem. Automation should be doing something the quickest and easiest way possible using as minimal effort as you can. if I’m trying to automate something small (the task itself with the amount of code you’re writing) and it goes over, say, 100 lines of code,  it’s probably not worth it. So 100 lines of code for something minuscule, I’d just walk away. It comes down to the size and value of what you’re automating, but I can see why not everyone wants to automate ‘everything’, people enjoy the satisfaction of pressing buttons and it being done, but the process should be as minimal as possible. 

Do you see automation taking over roles in the future?

Not in the next year, or the next 2 years but at the pace tech and automation is moving roles may be changed because automation can now do what once required a techy. I’ve seen in a short time how quickly things move, it’s scary to think how fast it’ll move in 10 years time. The more people start trusting programmes and software, the more we’ll see roles shift to accommodate this. 

If you’re interested in learning more, take a look at our services page.