Is it time to kill staging?

Some people argue that it is.
"Just test in production."

Before you grab the pitchforks, here's their argument 👇.

We spend time, budget, and effort maintaining a "production-like" environment...
that rarely behaves like production because it has:
- Different data volume
- Different IAM / permissions
- Different traffic & load
- Different treatment, production is sacred, staging is a dumping ground

The proof? Just count how many bugs pass staging and blow up in production.

Production is better for
✅ Monitoring
✅ Performance
✅ Real data edge cases
✅ IAM issues
✅ Scaling and load

Conclusion? Let’s kill staging and test in production.

When asked, “Isn’t that dangerous?”, the answer is:
“It’s fine with feature flags, progressive rollouts, instant rollback plans, and strong monitoring.”

Here's our take: If staging isn’t working… something in the process needs improvement.
The question isn't
Should we kill staging?
It's
How can we fix our staging?

Because the argument for killing staging ignores what people actually use staging for:
Staging is where we:
✅ Run functional checks
✅ Catch logical bugs
✅ Review UX/UI
✅ Demo to clients

And many of the “production-only” tests can be done on staging:
✔️ Performance (load tools exist, some even use production logs)
✔️ Edge-case data (factories + synthetic datasets; or anonymised production data where allowed)
✔️ IAM scenarios
✔️ Scaling simulations

Monitoring? Yeah, we agree, that is a production concern.

Testing in production safely is possible, but in reality, it’s often more complex and higher risk than simply maintaining a good staging environment. If your only motivation is to eliminate staging, it’s probably not worth the trade-off.

The future isn’t killing staging, it’s better staging.

So let’s not take staging behind the shed, Old Yeller style, just yet.
Let’s evolve how we test, not abandon what still works.

Staging still has a place in the development workflow; it’s one of the best, simplest safety nets you can have.