Cloud costs can easily spiral out of control without careful management, although it is easy to reduce cloud costs by not paying for resources that are not being used. Stop idle resources and the cost savings can be considerable.
One of the biggest areas of cloud wastage is On Demand resources used for non-production purposes. Resources are spun up for staging, QA, and development. These resources are not needed 24/7/365, yet companies often leave these resources running. Cloud companies still charge for these resources even when they are idle.
Even if resources are needed 12-hours a day, that means for 50% of each business day the resources are idle and costing companies money unnecessarily. Turn them off overnight and the savings are 50% per resource. Turn them off over the weekend as well and the savings rise to 65%. Most companies that carefully control their cloud costs do just that. They schedule their non-production resources so they are not running when they are not being used.
With the average instance costing around $220 per month, a 65% saving is considerable – $143 can be saved on each instance each month. In the grand scheme of things that may not be much, but it certainly is if you are running 10 instances or 100 instances. That corresponds to $17,160 and $171,600 respectively per year.
ParkMyCloud, the leading cloud automation and scheduling platform provider, has highlighted four main areas where companies can achieve savings, simply by stopping paying for idle cloud resources. The company’s platform allows costs to be saved across the four main cloud service providers: AWS, Azure, Alibaba, and Google Cloud.
The main area for wastage are virtual machines and On Demand instances and their associated scale groups. These are commonly turned on 24/7/365 for no good reason, especially when these resources are used for non-production purposes.
Three other types of resources which are left running and accruing costs unnecessary are relational databases, load balancers, and containers.
Relational databases are often left running when they are idle. It is not always possible to turn off these resources, but certain types of RDS instances can be turned off on AWS. Unfortunately, this is not possible on Azure’s SQL Database or the Google Cloud Platforms SQL, but it is possible to rightsize these resources to avoid overspending.
Load Balancers – AWS Elastic Load Balancers (ELB), Azure Load Balancer and GCP Load Balancers – cannot be stopped when they are not used, but it is possible to remove them when there are no instances attached to them. Make sure you sent up alerts in Cloudwatch/Azure Metrics/Google Stackdriver and monitor those alerts to ensure you are not paying for what you do not need.
Containers are also a big area of waste, and while it is not easy to turn them on and off, at least not yet, you should be monitoring your containers and the utilization of the infrastructure especially in your non-production environments to avoid unnecessary overspending.
With a little management – and a suitable SaaS platform – it is possible to substantially reduce cloud costs without affecting performance. Other key areas where money is being wasted in the cloud is on overprovisioned resources. You can read more about commonly overprovisioned cloud resources here.