Elastic scaling method based on Kubernetes cluster

Abstract

The invention discloses an elastic scaling method based on a Kubernetes cluster, is applied to the technical field of computers, and aims to overcome the defects of an elastic scaling strategy built in Kubernetes, predicts future workload according to historical index values of the cluster on the basis of a deep learning platform. The absolute value of the difference between the current index value and the predicted index value is compared with the set threshold value, so that the condition of frequent capacity expansion and shrinkage of the service in the cluster can be avoided; if the absolute value is larger than or equal to a set threshold value, it is considered that an index peak or an index valley is about to occur in the future, the service in the cluster is necessary to execute capacity expansion or capacity shrinkage operation, and meanwhile, a timer is further arranged for recording time consumption of one time of capacity expansion or capacity shrinkage operation; in order to avoid the conflict between the current capacity expansion and contraction operation and the previous capacity expansion and contraction operation, a time threshold value is also set, and the capacity expansion or contraction operation is carried out only when the value of the timer is greater than or equal to the time threshold value.

Description

technical field [0001] The invention belongs to the technical field of computers, and in particular relates to a container elastic scaling technology. Background technique [0002] In a Kubernetes cluster, elastic scaling is mainly used to solve two unexpected problems in engineering systems: one problem is that due to excessive system traffic and increased system load in the short term, it is necessary to expand the specified service to ensure the stability of the system The other problem is that due to the low utilization rate of various resources in the system, the specified service needs to be scaled down, thereby improving the utilization rate of system resources and ensuring the rational use of system resources. [0003] The built-in elastic scaling policy of Kubernetes has the following shortcomings: [0004] 1. The scaling strategy at the node level has a strong dependence on the underlying physical cluster or cloud platform. If the number of nodes in the physical c...

AI Technical Summary

Problems solved by technology

[0004] 1. The scaling strategy at the node level is highly dependent on the underlying physical cluster or cloud platform. If the number of nodes in the physical cluster is too small, capacity expansion cannot be performed. However, the functions of the cloud platform and the settings and restrictions made by the cloud vendor related, resulting in poor versatility and limitations of these scaling strategies

[0005] 2. When implementing the service-level vertical scaling strategy, it is usually necessary to stop the operation of the original Pod first, and create a new Pod to carry the service according to the requirements of resource allocation. This process will cause the service to be interrupted for a certain period of time during operation.

[0006] 3. There is a certain lag in the responsive scaling strategy, which may lead to longer user request time and lower user experience

Images