Container cloud hybrid elastic scaling method based on resource demand prediction

A container cloud resource elastic scaling technology provides resource expandability for a micro-service architecture cloud application in a cloud computing environment, but dynamic fluctuation of an external load causes increase of response delay of the cloud application and enormous resource waste. Therefore, the invention provides a container cloud hybrid elastic scaling method based on resource demand prediction. A historical resource use data set is constructed through cloud application resource operation monitoring, a novel deep learning model adopting a plurality of convolution kernels and two-sequence information interaction is designed to accurately predict multi-time-period resource requirements of dynamic changes, and then expected resource quantities of horizontal and vertical elastic expansion are calculated respectively. And horizontal and vertical resource elastic expansion and contraction are combined to realize hybrid resource adjustment of the container cloud. According to the method, the r