Why I'm Not an SRE

SRE (Site Reliability Engineer) as coined by Google in 2003, came into prominence in 2016 with the release of “The SRE Book”. It defines a particular job in an engineering org. This job is mostly focused on of course, reliability, but also performance, observability/monitoring, code deployment, capacity planning, and many other adjacent activities.

Some people say SRE is Google’s “version of DevOps” (aka a “DevOps implementation”). This is a little tricky to talk about, because DevOps, started as a rallying cry for shortening deployment times and adding more automation in the SDLC, has turned into a role itself. There are lots of job openings for “Devops Engineer”, and engineers really identify that term, some calling themselves “a devop”.

While I don’t like the “devop” term, for the purpose of this blog post, SRE as an implementation of the DevOps philosophy is close enough.

The Stories We Tell Ourselves

Our jobs titles can strongly influence our personal identities. Think of your job title as a mini prompt fed to your own personal neural network, like:

I’m a good SRE. I build reliable systems. I’m good enough, I’m smart enough, and doggone it, people like me.

Why I Don’t Like SRE

To me, reliability has never been the top priority for most of the systems I’ve managed. It just hasn’t. Certainly reliability is one important aspect of a good system, but only to a degree.

For any individual job, there may be a time that reliability is job 1, but other times it might be performance, or cost efficiency. There are lots of tradeoffs to be made like:

  • For security, we need all logins to funnel through this central server
  • For efficiency, we need to run in 2 availability zones instead of 3.
  • For performance, we need to run the database as a singleton instead of being distributed.

I just don’t want to be boxed in, as simple as that. For that reason, I don’t identify as an SRE, and I just stick to “engineer” (luckily not a protected title in the US). That way I don’t let a reliability bias hold me back from making hard tradeoffs when the time comes.

Comment via email