Intro

Recently I heard that someone asked in an “IT Systems Engineer” interview a question: “Why are you using Agile in your team? Are not most of the tasks on your team ad-hoc?”. This made me think about this topic deeply. The work being done in these types of teams might seem distant from regular programming at the beginning. However, it is not because in general it is focused on automating stuff using software, avoiding manual labor, building reliable systems or, in other words, the product and your clients are just different – they are internal whereas usually, they are external. Let me try to explain.

Difference between sysadmins and SREs

Let’s first note the difference between those two – this is important because a lot of people have a wrong perception that everyone other than developers does not create any kind of software. Alas, the SRE/DevOps movement is mainly all about eliminating toil. For example, Google does not allow their SREs to spend more than 50% of their time doing manual stuff. That means that it should be done automatically by some kind of automation. On the other hand, the sysadmin model is about dividing the IT & development into two separate silos – the developers and the system administrators. Everyone is doing their own stuff and barely collaborating. Also, they are usually not creating much software of their own – mostly just small scripts with Perl or Bash. As you can see there is a stark difference in the mindsets and we are talking about the former one.

Iterative programming methodologies for SREs

Just like in the “normal” programming world, requirements change and it is an inevitable part of the whole gig, I think, unless you are making software for things where reliability is of utmost importance – for example, you would not want a plane to crash because of an overflow error (that still happens sometimes like in this case of Boeing Dreamliners).

The benefits and downsides of each programming methodology do not differ at all in any case. Let me propose an example. For instance, after a month your DNS service which you maintain might get a new requirement – the self-service should get a new batch creation feature. It would let create many records with only pressing a few buttons, saving the precious users’ time. This might not be apparent at the beginning.

This is where iterative programming techniques such as agile are useful because they solve this problem of uncertainty – they embrace it, and they let you pivot in the middle (between sprints) of your development process. With sequential development methodologies, you would have to wait until the end a whole cycle to implement any kind of new requirements.

Practically employing iterative programming in the day to day life of a site reliability engineer

For special, toil type of tasks, you could create one huge task in your time tracking software. For example, create a task in Jira, under which all of the other tasks would be created as sub-tasks where all of the nitty-gritty details would be written down and the time tracked.

Afterward, you could sum up the time spent on toil type of tasks versus the other ones. Then you could tell if you yourself or the engineers spend more or less than 50% of their time on this type of work which could subsequently become a signal that something is wrong.