Most cloud bills carry 20–40% waste. These ten changes are low-risk and can often be done in a single sprint.
Cloud spend tends to creep up quietly as teams ship features. The good news is that much of the waste is easy to recover.
Quick wins
- Turn off non-production environments outside working hours.
- Right-size instances using actual CPU and memory utilisation data.
- Delete orphaned resources such as unattached disks, old snapshots and idle load balancers.
- Use storage tiers: move infrequently accessed data to cheaper classes automatically.
- Commit to steady workloads with reserved instances or savings plans.
- Use spot capacity for fault-tolerant batch jobs and CI runners.
- Set log retention policies so you aren't paying to store everything forever.
- Review data transfer costs, especially cross-region and NAT gateway traffic.
- Tag everything so every cost has an owner.
- Set budgets and alerts so surprises are caught in days, not at month-end.
Make it a habit
The biggest savings come from making cost visible to the teams who create it. A monthly cost review alongside your reliability metrics keeps both in balance.