I built an AWS waste scanner, but I realized detecting waste isn't the hard part

I built a small tool called CloudSaver.

It's a browser-based AWS waste scanner. No AWS credentials, no automatic changes.

At first, I thought the hard part would be finding waste.

That sounded straightforward:

Find unused EBS volumes.
Find old snapshots.
Find idle instances.
Put them on a list.

But while working on it, I changed my mind.

Finding something that looks wasteful is often the easy part.

The hard part is the next question:

Can I safely delete it?​

Some patterns are easy to imagine:

• a snapshot that looks expensive but is actually an important rollback point

• a data volume that seems orphaned but is still referenced somewhere

• an instance that looks idle, but shutting it down breaks a recovery or automation workflow

And there is another subtle one with EC2: stopping an instance stops instance-usage charges, but attached EBS storage and other billable resources can remain. So even the phrase "this instance is costing us money" can be less simple than it sounds.

The more I worked on this, the more I felt that the valuable part wasn't just detecting waste.

It was making the decision safer.

That's why I ended up taking a fairly conservative approach with CloudSaver:

It doesn't automatically delete anything.

It gives you a list of things worth looking at, and the goal is to make the risks and checks easier to understand.

I'm still figuring out how useful that is in the real world.

Curious how other indie hackers / small software teams handle this:

When you know some cloud resources are probably unnecessary, how do you decide what is actually safe to remove?​

Jerry Chen