How Much of Your AWS Runs While Nobody Is Working?

How much of your infrastructure actually runs while nobody is using it?

For most teams I’ve seen, it looks something like this:

• Working hours: ~8 hours
• Still running, nobody using it: ~16 hours

Dev environments, staging servers, test databases — running twice as long as they’re actually used. And that’s just weekdays. Add weekends and the numbers get embarrassing.

It’s not negligence, it’s friction

Nobody decides to pay for idle servers. It just quietly happens.

Shutting things down manually means remembering to do it, knowing which instances are safe to stop, and dealing with the morning where someone starts work before everything is back up. So people leave things running. The cost is invisible enough that it never becomes urgent.

The actual numbers

If your dev environment runs 24/7 but you only use it 8 hours a day, you’re paying roughly 3x what you need to. Factor in weekends and it’s closer to 4x.

For a small environment that’s a few hundred dollars a month. For a bigger one, thousands. Same machines, same compute — just sitting idle for most of their lives.

Why I started building ParkMyAWS

That pattern is what pushed me to build ParkMyAWS.

It connects to your AWS account, you set your working hours, and it stops EC2 and RDS instances overnight and starts them again in the morning. That’s it.

You don’t need a cost optimisation strategy or a FinOps consultant. You just need to stop paying for hours when nobody is logged in.