No one really understands what we do. Not my family, not my friends, sometimes not even my coworkers. “DevOps” is this vague, catch-all label people throw around when they don’t want to think about what actually happens after code is written.
Here’s what it really means: I’m the person who worries when you don’t. I’m the one staring at dashboards at midnight when you’re asleep. I’m the one holding the pager when something random and catastrophic happens in production, and no one else can figure out why.
At my startup — which loves calling itself “cloud-native” and “blazing fast” — that means dealing with endless Kubernetes upgrades, broken Helm charts, flaky CI pipelines, and cloud costs spiraling out of control because someone left a test cluster running for three weeks.
I’m tired. Not because I hate the work — I actually love automation. I love the feeling of taking some fragile manual process and turning it into a clean, repeatable pipeline. I love making things resilient. But I’m tired of the hero culture.
Here’s the dirty little secret of DevOps: the better you are at your job, the less visible your work becomes. No one notices when things don’t break. No one writes thank-you notes when uptime is 99.99%. But the one time you miss an alert, or a deploy goes sideways? Everyone notices.
It’s exhausting. And honestly, the constant grind of on-call wears you down. No amount of “unlimited PTO” really compensates for that moment of dread when your phone buzzes at 2:14am and you know you’ll be debugging some weird race condition in a container that shouldn’t even exist.
Some days, I think about quitting. Going back to a pure developer role. Or finding a cushy job with no on-call, where no one expects me to be a silent guardian for their fragile stack. But then I remember: this is the part of tech that feels real. Infrastructure isn’t shiny. It’s messy, human, full of edge cases. You can’t fake it. You either keep the system running — or you don’t.
So for now, I stay. I automate what I can. I mentor junior engineers. I push for sane incident response policies. And when that pager goes off… I answer. Like always.