Before recommending any architectural change, we run a three-part audit: idle compute, oversized databases, and untagged/unowned resources. In every engagement so far, those three categories alone have accounted for 20-30% of avoidable cloud spend.
Idle compute is the easiest win
Staging environments running 24/7, autoscaling groups with a minimum size that's never actually needed, and dev clusters nobody remembers to tear down are the fastest money to recover. Scheduled scale-to-zero for non-production environments alone typically saves 10-15% of total compute spend with zero risk to production.
Right-size before you re-architect
Teams often jump to a bigger architectural change — moving to serverless, switching cloud providers — before checking whether their existing instances and databases are simply oversized for the load they carry. Pull 30 days of CPU, memory, and I/O metrics before making that call; it's common to find databases provisioned for peak-of-peak load running at 8% average utilization.
- Scale non-production environments to zero outside business hours
- Right-size instances against 30-day utilization, not launch-day guesses
- Tag every resource with an owner and a cost center — untagged spend is unaccountable spend
- Set budget alerts before the invoice, not after
The cheapest infrastructure change is turning off the thing nobody's using.
How much can a FinOps audit typically save on cloud costs?
In our engagements, addressing idle compute, oversized databases, and untagged resources has consistently recovered 20-30% of cloud spend before any architectural changes are made.
Should I re-architect before cutting cloud costs?
No — right-size and eliminate waste first. Architectural changes like moving to serverless are a second-stage optimization, not a starting point, and are often unnecessary once the obvious waste is gone.
Hayden Jackson
Get engineering notes in your inbox
Practical AI, cloud, and security delivery notes - no fluff, no sales spam.