I heard many people reffering to as "Cloud cost optimization needs to shift left", Today there are 3 ways to provision capacity on cloud (think about it 3 legs of stool) 1. On-demand instances offer immediate access to compute resources they are often associated with cost inefficiencies and unpredictable expenses and can cause costs to spiral out of control if resources aren't managed properly. 2. Spot instances provide cost savings but can also be latency issues, and instances may be reclaimed by the cloud provider without prior notification, leading to disruptions. 3. Capacity reservations While it guarantees resource availability, there is a risk of underutilization. If resources are not actively used, organizations may still incur high costs, making this solution potentially expensive when workloads are not running at capacity." Flex Compute (4th leg of the stool), customized for Non-steady state workloads recently I recorded a podcast with CloudIDR team, What are some of the current limitations in cloud maturity and how this contributes to performance bottlenecks and potentially lead to unmonitored cloud spikes? Overall thesis is Flex compute approach, capacity is allocated only when needed and released as soon as it’s no longer required. This ensures cost efficiency without requiring any additional tools or creating a burden for teams. Link to the podcast I hosted earlier this week: https://medium.com/@saimsafder14/441c2eab0b7d

What’s the difference between On-demand instances and Flex compute? I see, So you’re reserving instances and rent it on-demand?

yes that's kind of true, but not fully sure of how, since I'm not part of their team, I did a chat with their C.E.O on my podcast he shed some light on it

@Saim Safdar I talk to Engineering leaders daily, and I am seeing a growing trend in driving cloud costs down, so cloud cost optimization is becoming increasingly important in today's rapidly changing cloud environments. The idea of shifting left in cloud cost management emphasizes the need for efficient resource utilization from the outset. By creating isolated testing environments within Kubernetes, organizations can significantly reduce infrastructure costs by eliminating the need for duplicate environments. This approach not only enhances testing efficiency but also supports cost-effective scaling, much like the benefits offered by Flex Compute for non-steady state workloads. Integrating these practices into CI/CD pipelines can lead to seamless testing and deployment processes, ensuring that cloud resources are used optimally and costs are kept in check.

You summed up nicely but just to extend a few points from business wrangle, we live in a world where software runs businesses and also removes/reduces toil, other businesses accept the cost as long as they're revenue-generating, but we have come along a way where every other business nowadays struggle with operational complexities, modern IT vendors either fall short of expectation or does that but they didn't have processes in place to keep track of their own cloud cost expenditure, there's a two-way traffic one vendors are fighting to keep their customers happy with cost equilibrium (until their business come revenue-able) and other they need to make sure they did that with what there IT budget except that right now I believe there's not much in the discussion that makes me scary, from the technical angle I believe without policies in place cost will eventually go high or un notice just to make customers happy all the time, I keep the technical angle short since you have said it perfectly.