AWS announced support for stopping and starting RDS instances for the first time. Along with the right monitoring and actions, this major release could help your teams save on RDS-type cloud database costs.
Over our years of serving enterprises and born-in-the-cloud companies, we’ve learned that two of the main cost drivers for the majority of our customers are from EC2 and RDS. Their instance hour costs contribute the most to monthly bills, according to our Usage Type analysis. Using a snapshot from our own analysis, RDS could account for as much as 19% of AWS spending in a given quarter depending on the size of customer and how many RDS instances are being used.
For every dollar you save on your complex cloud, it’s a dollar you can use elsewhere to drive more value for your business.
Get Started to Save More on Cloud Database Costs
Now that RDS instances can be turned on and off with a window of time, while still retaining their configuration, data and security policies, savvy folks with the right cost optimization tools at hand can improve RDS elasticity to save. It’s another great option for saving big in the cloud and can be used in addition to rightsizing resources as part of an RDS optimization plan.
Making RDS as Elastic as EC2
Prior to this news, stopping and starting instances was only done with EC2. Our customers utilize EC2 instance by turning off instances during “nights or weekends,” or any known downtime, in order to reduce costs.
Until the AWS announcement, RDS could not be easily turned on and off, meaning instances across development, test and production environments were running 24/7 even if not required, driving unnecessary costs. Or, there would be a bit of extra work to spin up the instance and reload data, incurring data transfer costs and other charges depending on the RDS setup.
Now RDS instances can be stopped and started, without being terminated. According to the AWS announcement, “Amazon does not delete any of your automatic backups or transaction logs…” allowing users to restore via point-in-time backups once they turn their RDS instances back on. This option is specifically available for RDS instances running in single-Availability Zone deployments that aren’t a part of a Read Replica Configuration. AWS invites curious users to visit their official documentation for more details.
Before you head over there, let’s take a look at a couple of examples using Cloudability to illustrate this cost savings advantages of this RDS announcement.
You can see that in this example, through a combination of auto-scaling and stopping EC2 instances during weekends or other “down time,” the number of instance hours and the cost for EC2 drops significantly during these periods. The savings during these pauses really add up.
Visualizing the Potential RDS Elasticity and Savings
Now let’s take a look at the breakdown of costs for RDS.
As you can see, the instance hours are consistent and flat, driving unnecessary costs even during off-peak hours. This is typical of many use cases prior to the AWS RDS announcement. With the ability to improve RDS elasticity, we can start to work in programmatic ways to turn RDS instances off during downtime, such as nights, weekends and other breaks contextual to your workloads.
Gain Huge Savings Potential from RDS Dev and Test Environments
While your storage and snapshot costs will persist, any customers with a healthy test and development spend with RDS can now look at stopping instances during off hours. This introduces elasticity into your hourly RDS usage, reducing those daily instance hours and your monthly bills.
For Cloudability users, you can start strategically planning your RDS elasticity now in the app.
Improving RDS Efficiency Opens New Reserved Instance Possibilities
Once your RDS instances are rightsized and you’ve introduced elasticity, it’s even more lucrative to invest in RIs that can lower cloud database costs. Cloudability’s RI Planner philosophy fits perfectly with this model, building a histogram of your hourly usage to give you precise recommendations for RI purchases. Lower rates from RIs are a great compliment to your elastic RDS workloads.
Stopping and starting RDS instances will be reflected via usage hours in both our rightsizing tool and the RI Planner by looking at your utilization threshold over the time period, recommending the appropriate RI purchase for your needs. To see this kind of analysis and recommendation in action, check it out in our app!