You may, like me, have become quite interested in Continuous Delivery and all of the associated ‘things’ - agile, devops, kanban but if, like me, you don’t work at the Google’s or Netflix’s of this world you may sometimes wonder “Where to start?”

That conundrum can be made more difficult to fathom out if you spend the majority of your working time on operational issues.

You become caught in a vicious cycle of wanting to improve things so you have less operational issues but the operational issues prevent you from devoting constructive time to improving things.

I used to get quite frustrated with this cycle because I would often learn of shiny new stuff that X was doing and think “why can’t I do that?” and the simple answer was “you are not X”. Once I had accepted that it became even clearer where I should start. As a nix *admin I needed to keep it simple, do one thing and do it well and plug that one thing into other one things that did one thing well.

😀

To learn more on the “Unix Philosophy” the Wikipedia page is a good start: Unix_philosophy

So where did I start?

![don’t fix all the things](/img/dont_fix_all_the_things.jpg# center)

I adopted another (heavily paraphrased) idea I came across a long time ago (but now cannot find a reference for):

So I continue to work on operational issues. I will help someone (usually devs) to progress their tasks. I will aim to improve one thing a week.

I’ve been falling short on my target of improving one thing but I know if I improve one thing a week I will free up more time to improve something more regularly.

So where did I actually start?

I migrated the infrastructure into critical and non-critical systems. Issues on non-critical systems are no longer service impacting.

I am working on bringing Config Management back in the form of Puppet. I chose Puppet because I know it’s syntax and it will get me off the ground quickly. Other config management tools are available. And any config management is infinitely better than none right?

Next I plan to improve the build pipelines and fix up the data dumps from production to dev.

I aim to continue this cycle of small improvements and hopefully I will end up with a lot more time to be able to do the more exciting stuff that I see others doing with envious eyes! And I’ll be blogging about the simpler steps that I hope will get me there.

Maybe you’re in a similar situation?