PruneOps launches in beta: real FinOps for AWS

Why PruneOps exists

Cloud bills grow silently. A resource gets provisioned for a launch, a test, a migration — and then stays running, forgotten, adding to the bill month after month. The usual way to catch this is opening Cost Explorer and hunting for patterns by hand: runaway growth, services eating up half the bill, things nobody has touched in months.

PruneOps automates exactly that work. Instead of a periodic manual review, it keeps a live inventory of everything deployed and, from there, flags where you're overpaying.

What it does

• Continuous inventory — discovers and syncs resources automatically, with no need for periodic manual reviews.
• Recommendations with numbers — resources to delete, others to right-size, with the estimated savings next to each one.
• Operational risk detection — configurations that don't cost money directly but do cost headaches, like open ports or overly permissive policies.
• Terraform comparison — checks live infrastructure against state files, to identify what's running that nobody manages from code.

All of this rolls up into a FinOps Health Score for compute, storage, databases, and network, with the possible improvements broken down per area.

Metrics over estimates

One of PruneOps' design principles is not settling for estimates. Recommendations are based on real usage metrics (CloudWatch), not generic heuristics applied the same way to everyone — that difference matters when the decision is stopping or resizing a production resource.

Current coverage and what's next

Right now PruneOps covers the main AWS services. The plan is to keep broadening that coverage progressively, and later reach other cloud providers.