Though RhinoDox’s sophisticated DevOps team knew of Savings Plans (SP) and Reserved Instances (RI) to reduce EC2 spend, they didn’t have the bandwidth to manually manage the SP and RI as they scaled up, and were concerned about making long-term commitments they couldn’t exit from.
Erratic AWS reservation pricing made estimating potential EC2 savings difficult. Figuring out which instances would benefit from the greater-flexibility but decreased-savings of Savings Plan and which ones would benefit from the greater-savings but decreased-flexibility of Reserved Instances was a complex calculus. Going all-in on a Savings Plan would mean the greatest flexibility but the least possible savings. Going all-in on Reserved Instances would mean the greatest savings but the least amount of flexibility. Trying to figure this out while managing a complex cloud infrastructure is not something the RhinoDox team wanted to do.
We’ve used other cloud spend monitoring tools before, and they were fine at making recommendations - this server is too big or you should buy a reserved instance here - but they don’t do the work for you so someone still has to take the time to go do that work, and ultimately those were really valuable early on but not so valuable as time moved on and we got a handle on things, whereas Usage is going to be there sort of being a little bit of a janitor for us as we go along and pointing out ‘hey, you added 5 servers you don’t have reserved instances for and if you buy RIs you can save X’ that’s a real value that it’s automatic.” – Travis, Head of Engineering, RhinoDox
Furthermore, both of these options required committing to Amazon for a minimum of 1 year and up to 3 years for the maximum savings. This meant that the RhinoDox team would have needed to forecast usage for the next 1-3 years to minimize excess spend on unused commitments. This adds even more complexity to the already complex AWS reservation pricing mode.
All of our customers grapple with the problem of figuring out the right blend of billing modes for the best possible savings and flexibility, and Usage.AI not only surfaces the right recommendations, but also applies the savings for the customer at the press of a button.
The RhinoDox team explored other tools like CloudHealth and found some initial value with them, but was disappointed with the manual effort required to get the savings. Tools like CloudHealth would show the savings, but the RhinoDox team would have needed to add cards to their sprint in order to get the savings implemented, which took them away from higher priority features.
When using Usage.AI, RhinoDox was able to see the possible savings on the Usage.AI dashboard and also get the savings by simply pressing ‘Approve’. Usage reduced RhinoDox’s EC2 spend by a whopping 39% by creating a blend of Savings Plans with Reserved Instances, all backed by Usage’s guaranteed buy-back. All the RhinoDox team had to do was press “approve” and the savings were instantly applied. That’s it, zero touch to the servers, no code change required, no engineering or R&D work needed, and zero downtime.