There are numerous Cloud Service Provider (CSP) FinOps products that can review, collate, summarize, and recommend ways to optimize your cloud spend. If you’re using one or more cloud providers and don’t actively manage your Cost and Usage Reports (CURs) on a daily basis, investing in such a product is a smart move. A great example is CloudForecast.io .
However, beyond using FinOps tools, there are fundamental steps you can take to ensure you save money. Here are my top five recommendations:
1. Turn Off Resources You Don’t Use
This applies especially to compute instances, including EC2, RDS, ElastiCache, SageMaker, and others. Any unused resources left running will add unnecessary costs. It may seem an obvious step, however even the smallest organization leaves unused resources on.
2. Schedule Resources That Are Only Needed Part of the Day
If your non-production environment is only used within U.S. business hours across four time zones, you can optimize your instance spend (e.g. RDS) to run for just 55 hours per week (11 hours × 5 days) instead of the full 168 hours per week. This alone can save you over 65%. Using CloudWatch metrics can help identify candidate resources for scheduling.
3. Avoid Over-Provisioning
This requires both reactive monitoring (tracking your cloud spend) and proactive prevention (setting up controls).
- Track resource creation to capture over-provisioning immediately, not a day later.
- Implement IAM policies to restrict instance provisioning with a whitelist of approved instance types.
- Enforce a least-privilege security model to prevent teams from accidentally launching oversized instances (e.g., a 24xlarge instance sitting idle).
4. Stay Up to Date With AWS Releases for Instant Savings
Simply keeping up with AWS release news can lead to instant savings.
- When AWS Graviton2 Gen 6 instances (r6g, etc.) were released in 2020, switching from Gen 5 resulted in a 10% cost reduction, with even greater savings for older generations.
- In 2024, AWS introduced Valkey , a wire-compatible, managed Redis replacement that provides a 20% savings over ElastiCache for Redis.
5. Consider Reserved Instances—But Optimize First
Reserved Instances (RIs) can significantly reduce costs, but only after you have exhausted all other optimization opportunities:
- Turn off non-production workloads when not in use
- Upgrade to more cost-efficient instance types
- Right-size existing resources
- Understand RI lock-in rules (e.g., account, region, instance family, and product type)
- Leverage size-flexible instance types that provide greater granularity, e.g. EC2 and RDS
Bonus Tip: Explore Alternatives to Managed AWS Services
AWS offers a vast range of managed services, but marketplace providers can sometimes offer more cost-effective, drop-in replacements with minimal changes. For Example:
- Deploy ScyllaDB instead of DynamoDB Read More
- Use ReadySet as a drop-in cache instead of modifying your app to use ElastiCache Read More
- Aiven’s MySQL for a more true open-source experience than AWS RDS MySQL Read More
There are many alternatives worth considering, and evaluating them could lead to substantial savings while maintaining or even improving performance.
