Virtual machine capacity planning

摘要

Virtual machine capacity planning techniques are disclosed. In various embodiments, a set of time series data is constructed based at least in part on virtual machine related metric values observed with respect to a virtual machine during a training period. The constructed time series data is used to build a forecast model for the virtual machine. The forecast model is used to forecast future values for one or more of the virtual machine related metrics. The forecasted future values are used to determine whether an alert condition is predicted to be met.

主权项

1. A method of virtual machine capacity planning, comprising:
constructing a set of time series data based at least in part on virtual machine related metric values observed with respect to a virtual machine during a training period; using the constructed time series data to build a forecast model for the virtual machine; using the forecast model to forecast future values for one or more of the virtual machine related metrics; and using the forecasted future values to determine whether an alert condition is predicted to be met, wherein using the forecasted future values to determine whether an alert condition is predicted to be met includes computing based at least in part on the forecasted future values a forecasted summary statistic and comparing the forecasted summary statistic to an alert threshold, wherein the forecasted summary statistic is computed at least in part by computing a Hotelling T2 distance between a forecast feature vector associated with a future period and a mean vector associated with the training period, wherein the alert threshold is based at least in part on a difference between an observed feature vector associated with the virtual machine and a mean vector associated with the virtual machine, and wherein a responsive action is taken based at least in part in response to determining the alert condition is predicted to be met.