Some call it penny-pinching, some call it watching my cash-flow, but everyone is trying to avoid the price creep that might happen if you are using the cloud. An honest mistake can lead to terrible effects – and if you don’t know your own architecture well enough, this can be disastrous. So, here are some strategies to effectively avoid creeping cloud hosting prices:

• Monitor usage and costs regularly: Use tools such as Azure Cost Analyzer, AWS Cost Explorer or Google Cloud’s billing dashboard to keep track of resource utilization and costs. This will help you identify areas where you can reduce spending, such as underutilized instances or overpriced storage services.
• Use reserved instances or committed use contracts: These provide a significant discount compared to on-demand pricing by committing to use a specific amount of resources for a set period of time, e.g., 1 or 3 years.
• Take advantage of spot instances: Spot instances are unused Azure or EC2 instances that are made available at a discount to users willing to take the risk of having their instances terminated when the spot price increases. Yes, it needs to think through your microservices infrastructure and your circuit breakers, but the savings can be tremendous.
• Optimize your infrastructure: Right-sizing instances, using auto-scaling groups, and using managed services like Azure Managed Cloud Database Service or AWS RDS instead of running your own database servers can help reduce costs.
• Use managed services: For me, this is probably the biggest pet peeve: Using managed services like Azure Functions, AWS Lambda or Google Cloud Functions instead of running your own servers can greatly reduce costs and eliminate the need for infrastructure management. Believe or not, you would not be able to optimize your custom solutions, let it be written by you or using some 3rd party solution or similar, to be as cost effective as it has been done by the actual CSP – they are interested in making it cost effective, to be able to handle a better compression ratio on the same hardware, it would be for sure be compatible with other services, authentication would work out of the box, etc.

You can use a combination of various tools to achieve this, just look at these examples:

• Dropbox: Dropbox was able to save over 30% on its cloud hosting costs by using reserved instances and spot instances to reduce its spending on compute resources. In addition, Dropbox optimized its infrastructure to make it more cost efficient, including reducing the number of instances it was using, right-sizing its instances, and using auto-scaling groups to ensure that its resources were being used optimally. This combination of strategies allowed Dropbox to reduce its overall cloud hosting costs while maintaining the same level of service and reliability.
• Capital One: Capital One reduced its cloud hosting costs by over 50% by using a combination of reserved instances, managed services, and a cost optimization program. Capital One adopted a proactive approach to cost optimization, monitoring its cloud usage and costs on a regular basis, and implementing cost optimization strategies when necessary. For example, Capital One used reserved instances to commit to a certain amount of resources over a set period of time, which allowed it to receive significant discounts compared to on-demand pricing. In addition, Capital One adopted managed services like AWS RDS and AWS Lambda to reduce the amount of infrastructure it needed to manage and maintain.
• Expedia: Expedia reduced its cloud hosting costs by 20% by using reserved instances, auto-scaling, and by optimizing its infrastructure for cost efficiency. Expedia adopted a multi-pronged approach to cost optimization, which included committing to a certain amount of resources over a set period of time through reserved instances, using auto-scaling to ensure that its resources were being used optimally, and right-sizing its instances to reduce the amount of resources it was using. These strategies allowed Expedia to reduce its cloud hosting costs while maintaining the same level of service and reliability.
• SoundCloud: SoundCloud reduced its cloud hosting costs by over 50% by moving from on-demand instances to reserved instances and by optimizing its infrastructure for cost efficiency. By using reserved instances, SoundCloud was able to commit to a certain amount of resources over a set period of time, which allowed it to receive significant discounts compared to on-demand pricing. In addition, SoundCloud optimized its infrastructure to reduce the amount of resources it was using and to ensure that its resources were being used optimally, which allowed it to further reduce its cloud hosting costs.
• SmugMug: SmugMug reduced its cloud hosting costs by over 60% by using reserved instances, spot instances, and by optimizing its infrastructure for cost efficiency. SmugMug adopted a multi-pronged approach to cost optimization, which included using reserved instances to commit to a certain amount of resources over a set period of time, using spot instances to take advantage of unused EC2 instances that were made available at a discount, and optimizing its infrastructure to reduce the amount of resources it was using and to ensure that its resources were being used optimally. These strategies allowed SmugMug to reduce its cloud hosting costs while maintaining the same level of service and reliability.
• Netflix: reduced its cloud hosting costs by over 80% through a combination of reserved instances, spot instances, and optimizing its infrastructure.

Of course, as always, it is important to note that these strategies and tools will vary based on the specific cloud provider and your use case, so it is important to carefully evaluate your options and choose the best solution for your needs. However, these case studies provide a good starting point and demonstrate the potential savings that can be achieved through a proactive approach to cloud hosting cost optimization.