Since releasing Cloud Advisor three months ago, we’ve been hard at work trying to understand how customers use the tool and what they want from it next. Today, we share news of our first wave of significant enhancements to date, which went live on February 23, 2021. The enhancements include an expanded set of fix-it flows to implement recommendations easily, the ability to customize and override recommendations, a new recommendation to terminate idle Compute instances, and tweaks to our recommendation to resize underutilized compute instances to simplify and improve recommendation accuracy.
Expanded Fix-it flows
Fix-it flows provide an easy path resolution for each recommendation through a quick UI workflow. Fix-it flows make it easy to bulk apply a recommendation across an entire group of resources. With the new release, Fix-it flows can now delete boot or block volumes, resize underutilized Compute instances or database instances, stop idle compute instances, and enable object lifecycle management on buckets. These flows make it easier than ever to groom your cloud and ensure that you’re operating waste-free. For details on each recommendation, visit Implementing Cloud Advisor Recommendations.
Recommendation customization and overrides
Workloads are not one-size-fits-all, so recommendations shouldn’t be either. With recommendation customization and overrides, customers can now use one version of a recommendation for their production environments and another version for development and test environments. To learn more about this topic and how to start taking advantage of overrides today, read Customizing Cloud Advisor.
Stop idle compute instances
We’ve discovered that customers often deploy virtual machines (VMs) for a point-in-time purpose, run a task, and forget to terminate the machines when they’re done. With the “Stop idle Compute instances” recommendation, customers who forget about VMs can quickly discover them and easily terminate them in bulk. Read more about this recommendation.
Downsize underutilized Compute instances
Having reviewed the data from our first three months of life, our team discovered that the recommendation was more aggressive than we wanted. To improve things, we’ve reduced the utilization threshold and are now evaluating network throughput, CPU utilization, memory, and supported VNICs. We’ve also fixed a major source of confusion for customers: recommendations showing implemented when no changes were made. This notification occurred when instances had non-standard consumption patterns. We’ve modified how implementation is considered, and now only instances that have been downsized are eligible to appear as implemented. Read more about this recommendation.
Conclusion
We’re proud to bring these enhancements to you and excited to see how you take advantage of them. Over the coming months, we continue to expand Advisor’s feature set and library of recommendations. Our current top priorities include recommendations to improve resource performance and availability and the ability to customize the window of time over which resources are evaluated.