Hey everyone! We're a B2B SaaS focusing on data analytics, and our Azure bill is hitting over $6000 a month. A large portion—about 50%—comes from our SQL server (800 DTU and 1 TB) and Azure Data Factory, which we use to connect to Salesforce. Is this distribution of costs typical or unusually high? Looking for insights and suggestions on how to manage these expenses better!
5 Answers
What service tier are you currently using for your SQL database? It could make a huge difference in managing costs.
What kind of Integration Runtime are you using with ADF? Make sure you're controlling the autoscaling features to avoid needless compute costs during activities. And definitely check out different SQL SKUs for better pricing.
We discovered that a big part of our bill was coming from Azure Data Factory, so we transitioned some of that workload back into .NET code by using Hangfire for scheduling and stored procedures instead. That really helped reduce our expenses.
Your total isn’t outrageous for a B2B SaaS doing analytics, but if half of it is on a single SQL DTU, you may be overpaying due to the 'DTU tax.' I recommend switching to the vCore model since that could be 25-40% cheaper. Also, check the number of Integration Runtime hours and Data Movement Units in ADF, as that can add up too. Optimizing your usage could save you a lot!
Absolutely! Also, consider moving to self-hosted IR on a small VM if you're looking for consistent cost savings.
If you're looking at a $6k bill, it might be worth considering switching away from the DTU model for SQL. Non-DTU SKUs typically offer better IOPS which could fit your needs with a smaller instance. Also, if you haven't already, explore reservations and savings plans to bring that cost down.
Great point about the vCore model! That was a game-changer for us. Also, consider adjusting the polling frequency with Salesforce to reduce costs.