Last week we pushed the button on the latest release of our CycleCloud software for managing cloud HPC and Big Compute workloads.
This release has one particular feature that many customers customers asked for: Cost Alerting. This new feature will give you the ability to easily set cost alerts on a per-cluster basis. You can set the alert to be dollars per day or per month. This gives you a great way to manage consumption and assure that users aren’t blowing through budgets. After all, you want to give your users access to unlimited compute, but you don’t want to give them an unlimited budget.
Clusters from any supported cloud service provider display an estimated compute cost along with the core-hour usage. Daily or monthly budgets are set from the cluster page and trigger alerts when the threshold is crossed. Because the appropriate action when a cluster goes over budget varies, the CycleCloud software does not take any automated enforcement action. We find most customers try to set it the threshold to some percentage of total budget to give them a heads up before exceeding the budget. The percentage can be a function of the type of work and size of budget.
In addition to the cost alerting, we’ve added additional features to our Microsoft Azure support. CycleCloud now uses Azure Managed Disks and Images for virtual machines, simplifying management of storage and improving performance. Azure instances will automatically use CycleCloud’s standalone DNS configuration to improve the experience for Open Grid Scheduler users.
Current customers can download CycleCloud 6.5.3 from the Cycle Computing Portal. If you’d like to learn more about how CycleCloud can enable your cloud HPC and big compute workloads, contact us.