Cloud optimization biz CAST AI says companies are still over-provisioning resources and paying too much as a result. It claims that in Kubernetes clusters with 50 or more processors, typically only 13 percent of the provisioned processors and 20 percent of the memory are used.
Kubernetes Pod Overprovisioning Demo
CAST is developing a platform to monitor the use of Kubernetes resources and compare them to what it estimates the workload actually requires. The figures in its latest study are based on an analysis of more than 4,000 clusters operated by customers before optimization.
The reason for this over-provisioning is largely due to excessive caution on the part of users, with DevOps teams keen to avoid running out of memory, and also due to the difficulty of knowing exactly how much resources might be needed from the outset. Users may also be confused by the wide selection available, with AWS offering 600 different EC2 instances, the study says.
CAST believes this can be seen in the percentage of unused memory that is virtually identical across the three major cloud platforms – AWS, Azure and Google Cloud – meaning it is not due to the specifics of any particular cloud.
