Saved 70% on Cloud Costs While Decreasing ETL Load Times
Using Envisor Cloud Control, Senturus was able to accurately determine the best Azure capacity level required to meet their ETL database processing needs. By dynamically scaling Azure to meet peak and low demand times at the optimal service level, they saw data import performance accelerate by 85% and overall costs decrease by 70%.
Senturus relies on 50 Azure SQL databases to address the needs of our consulting, training and software development businesses. Like most organizations running a cloud environment, we were paying for capacity we were not using. We had sized our Azure resources to accommodate peak demand, but that capacity remained idle for large portions of the day.
The situation is well illustrated by one of our ETL databases. We would run daily data imports from external web apps into an Azure SQL database. To meet delays in the data import process, our admins responded in the typical manner, bumping up the Azure SQL database performance level to speed up the data imports. The increased service level solved the performance issue but wasted money as the database would remain idle for rest of the day.
We wanted a way to intelligently monitor and manage our cloud capacity and avoid paying for unused capacity.
We developed Envisor Cloud Control. With it, we are now able to accurately understand our server utilization, match it to the optimum required capacity level and dynamically scale the server to meet demand.
The Azure SQL database was running at an S2 level for the entire 24-hour period. We found that the database was under-provisioned for a couple of hours early each morning during the ETL load and idle (over-provisioned) the rest of the day.
With this insight, we determined that moving to the S4 tier, while the most expensive level, would be the most cost-effective. We then used Envisor Cloud Control to scale up the Azure SQL database during the ETL loads when we need more performance, and back down when we only need it to be generally available.
New resource scaling schedule
Envisor Cloud Control allowed us to rethink our computational performance needs and accurately manage our Azure capacity. Moving to the S4 tier gives us the computing power we need to meet peak demand and also decreases the load time.
For this one ETL database, we have seen impressive cost savings and performance improvements:
- Data import performance accelerated by 85%.
- Data is delivered to users 90 minutes earlier than previously.
- Overall costs decreased by 70%.
Envisor Cloud Control can reduce costs and improve performance for other cloud resources as well. Read more about Envisor Cloud Control.
Using Envisor Cloud Control, we saw data import performance accelerate by 85% and overall costs decrease by 70%.