Gallery of Savings
Below are a few samples of applications managed by AutoScalr along with screenshots of the savings being generated.
Application: First stage of processing for SessionCam's advanced web analytics SaaS offering.
Challenge: Application averaged between 100 - 200 instances and was growing linearly with the growth of their customer base. They wanted a way to reduce the growing cost but maintain the availability requirements since the application is core to their business offering.
Solution: AutoScalr allowed SessionCam to power the cluster using a blend of spot instances and lower risk instance types to save over $10,000 monthly while still meeting the stringent availability requirements.
Application: Queue-based back-end processing of customer data
Challenge: During heavy holiday season, some instances were detected becoming inefficient in processing after about 8 hours, causing the size of the cluster to scale up to handle the load. The root cause was believed to be a subtle memory leak.
Solution: Activate AutoScalr's Time-To-Live feature and set it to automatically replace instances in the cluster after they had been running for 4 hours. This effectively prevented the problem from occurring and allowed staff to focus on higher-priority items during the holiday season and address the memory leak issue during the next scheduled update for the component.
Application: SaaS Analytics Service for iOS Mobile Application Developers
Challenge: AgileMobile's service required heavy back-end processing running in an ECS cluster that needed to scale quickly based upon customer usage. The required tasks had different CPU/Memory requirements which complicated picking the most cost effective instance type to use. The end result was significant EC2 costs per customer and hard to manage autoscaling rules. They looked at Fargate as a way to simplify but concluded the per customer cost would actually go up with Fargate.
Solution: Implementing AutoScalr reduced operational cost by 65% and eliminated the need to manage the EC2 instances in the ECS cluster while automatically selecting the most cost effective instance type (CPU vs Memory) based on the currently running set of tasks.
Application: Back-end for leading conference dialer app
Challenge: iTapLess's leading conference dialer app had growing back-end EC2 costs to process the 10's of thousands of conference calls started daily. They wanted to reduce costs without a lengthly development project to achieve.
Solution: Implementing AutoScalr took less than 15 minutes and reduced their costs by 65%.