Contents
Nothing is fixed. It depends when and where.
I’ve watched DevOps engineers hate on Terraform and I’ve watched others treat it like scripture.
Debating an IaC tool without context is pointless—like arguing whether coffee is good based on someone else’s hangover.
Still, a Reddit thread on clickops vs IaC nudged me to write down how I think about Terraform as an ops person right now.
I first touched Terraform on an AWS project in 2019. Declaring infra in code felt like cheating in the best way.
I was mostly dev then, but everything after that gravitated toward Terraform—Aliyun, GCP, Azure, the whole tour.
Before the nice parts, here’s what I keep hearing from people actually running platforms.
“Just click it in the console”
Clickops is still the default everywhere. Console UI, done.
You want me to learn HCL, wire CI, and babysit modules for the same outcome?
Even if I bite that bullet—who gets the time back? How much of that work reuses on the next project?
“Efficiency” is a hard sell when the saved hours aren’t mine. Fair point.
“The cloud’s Terraform story is half-baked”
Don’t tell me you’ve never tried IaC on a newer cloud whose API feels like a first draft. Is the provider team even on speaking terms with the product team?
You start proud with resources in code. Provider bumps a version and suddenly you’re debugging upstream, opening tickets, watching issue threads.
Half the resource types aren’t in the provider anyway. So you maintain a runbook: what’s code, what’s manual, what’s “manual until the provider catches up”—and the docs lag reality by a quarter.
Some clouds at least give you an escape hatch (REST API provider, generic whatever). Others? Good luck.
“The pile never shrinks”
I’ve run pipelines with thousands of resources. Modules let app teams add more. They rarely volunteer to remove stuff.
The state file keeps growing. Plans get slower. Splitting state helps a symptom, not the culture.
Modules are work. Provider bugs become multi-resource incidents. Terraform didn’t magically make governance someone else’s problem.
“I’m the only one using it”
Worst case: you’re all-in on Terraform; everyone else clicks or runs one-off scripts they like better.
You preach modules. They nod. They keep doing what works for them.
Drift is the punchline. Someone edits “your” resource in the console. State lies. You find out in a 2 a.m. diff. Can’t really blame only them—you both live in the same account.
“Why not let AI click for me?”
We’re in the era where a model can spit out Terraform—or call the cloud APIs and CLIs directly.
Feels instant. No console, sometimes no pipeline.
Worried it’ll go rogue? Have it write a small script instead. Still automation.
Platform-wide reuse, guardrails, audit trail—“we’ll add that when we need it.” Maybe. Most teams aren’t there yet.
Side note: Predictive text will embarrass you in any language. I once fat-fingered an idiom in a draft title and missed it until the second read. (Guilty.)
If any of this sounds familiar, you already know the punchline.
Terraform isn’t the problem. Whether you have the org conditions to run it—and whether anyone will keep governing it—is.
If you’re still happily on Terraform today, honestly, congrats. That’s a decent place to be.
Questions? Leave a comment below or email: newbvirgil@gmail.com![]()
