Here are some generic cost optimization tricks that work for most customers. Try them now!!!

• Review the Elastic Compute Cloud (EC2) section. Are they using previous-generation instances? Upgrade them to the latest generation. Can their workloads support the Graviton2-powered instances? If yes, change the instance type to save costs with similar or better performance. Use the EC2 Instance Explorer to find the correct instance types, or run Compute Optimizer
• Under the EC2 section, check the cost of NatGateway. If it is significant, consider removing NAT Gateway from non-critical private subnets and instead use a t3.nano NAT Instance. It is less highly available but at a much lower cost, and for non-critical workloads for many customers it can be a useful compromise, especially if they only use NAT for infrequent outbound requests for software updates, etc
• Check the EBS volumes. If they are using gp2 volumes, recommend upgrading to gp3 volumes which provide better performance at ~20% lower cost and can be upgraded in-place.
• Check the EBS volumes section to see if they have provisioned IOPS and io1 volumes. Most workloads that previously needed io1 volumes for provisioned IOPS can now run on gp3 volumes with IOPS provisioned separately from storage.
• Check how much are you spending on EBS Snapshots. If it is significant, go to the EBS section on the console and review the list of snapshots. Are there duplicate snapshots, or very old snapshots that the customer does not need anymore? Delete them when appropriate. Some third-party tools create multiple snapshots that you may be unaware of.
• While on the EBS console, review the list of EBS volumes. AWS Console Cost-saving techniques that are effective for the majority of clients? Review those, take a snapshot if needed, and delete them.
• Back to the bill, review the Simple Storage Service (S3) section. Are you using only S3-Standard tier which has a significant amount of storage (several TB)? Recommend either turning on Intelligent Tiering or manually adding lifecycle rules to transition lesser-used objects to Infrequent Access and Glacier
• Check the Elastic Load Balancing section. If there are several ALBs running, review is the usage pattern requires them all. Sometimes, every developer will provision a new ALB for their workload, but using Host-based routing and Path-based routing, it is possible to consolidate traffic going to several targets into a single ALB, thus saving hourly costs of running multiple ALBs
• Review the DynamoDB section if there is a significant cost on it. Are you using On-Demand Read and Write capacity? If their usage pattern is predictable, moving to Provisioned Capacity can be cheaper
• Review the SageMaker section. See if the SageMaker RunInstance has a high cost. This can happen when someone provisions a SageMaker instance for testing something and forgets to shut it down at the end. Since SageMaker instances tend to be large powerful ones, cost adds up fast.
• Check with Data Transfer section. Is there a large regional data transfer cost? Typically listed at the end under Bandwidth: $0.010 per GB – regional data transfer – in/out/between EC2 AZs or using elastic IPs or ELB. If data is being transferred between VPCs in different regions, it should be set up to go over VPC Peering instead of via the internet, since it costs ~75% less
• Check if you are using EC2 Spot instances, Spot instances are ideal for workloads that are spiky, stateless, fault-tolerant, time-flexible, CI/CD, Machine Learning, pre-production, batch, container-based, HPC, web-applications, etc.

• Use Reserved Instances (RIs): RIs are a commitment to use a certain amount of compute or memory capacity for a certain period of time. In return, you get a significant discount on the on-demand price.
• Use Savings Plans: Savings Plans are a commitment to use a certain amount of compute or memory capacity over a one- or three-year period. In return, you get a discount on the on-demand price. Savings Plans are a good option for workloads that have predictable usage patterns.
• Use the AWS Free Tier: The AWS Free Tier offers a limited amount of free usage of many AWS services. This can be a good way to get started with AWS without having to pay any upfront costs.
• Use Volume Discounts: Volume Discounts are available for EBS volumes that are provisioned for a minimum amount of time. The longer the commitment period, the greater the discount.
• Stop or Downsize Unnecessary Instances: If you are not using an EC2 instance, you can stop it to avoid incurring charges. You can also downsize an instance to a smaller size if you do not need the same amount of compute or memory capacity.
• Delete Unattached EBS Volumes: If an EBS volume is not attached to an EC2 instance, you can delete it to avoid incurring charges.
• Analyze Your Usage: Use the AWS Cost Explorer to analyze your AWS usage. This can help you identify opportunities to reduce your costs.
• Use AWS Trusted Advisor: AWS Trusted Advisor provides recommendations for improving your AWS costs and performance.