Google Cloud Platform for Architects
上QQ阅读APP看书,第一时间看更新

Rightsizing recommendations

In the on-premise world, it pays to over-provision. If you need a server of capacity X today, you're best off buying one with a capacity of 1.5X—if only to not have to deal with finance and procurement all over again in a year's time.

When you switch to the cloud though, if you don't adjust that mindset, you might end up with sticker shock. On the cloud, remember that you can scale both up and down.

That's where the recommendations from GCE come in handy. Compute Engine provides machine type recommendations to optimize the resource utilization and minimize expenses of your virtual machine instances. These recommendations are generated automatically based on system metrics such as the CPU and memory utilization, gathered by the Google Stackdriver, which is a suite of GCP tools for Logging and Monitoring services over the last eight days.

You can then use these recommendations to resize your instance's machine type to more efficiently use a machine type's resources.