I'm seeing more SRE roles that look exactly like Backend Developer roles but with an on-call rotation. The original Google SRE book talked about a 50/50 split between dev and ops work, but it feels like the "Ops" part is disappearing in favor of complex Go or Python development for custom tooling. Is the "Systems" knowledge part of SRE becoming less important?
3 answers
Systems knowledge is actually more important than ever because of how complex distributed systems have become. While I spend 60% of my time writing Go code for our internal controllers, that code is useless if I don't understand TCP/IP, kernel tuning, and how our service mesh handles traffic. The "coding" is just the medium we use to express our systems knowledge at scale. You aren't just a dev; you are an engineer who uses code to solve operational problems that can't be fixed by hand anymore.
Do you think the high barrier to entry for coding is scaring away talented traditional SysAdmins who have deep infrastructure knowledge?
At the end of the day, SRE is about reliability. If you can ensure 99.99% uptime with bash scripts or Go, the tool matters less than the result.
Exactly, Rachel. Though I will say, maintaining a thousand bash scripts is a nightmare compared to a well-structured Go project!
Kevin, I definitely think there is a "skills gap" there. Many great admins are struggling with the transition to full-scale software engineering. However, I’ve found that mentoring them on basic automation scripts usually acts as a bridge. We need their "disaster recovery" mindset even if they aren't the best coders yet.