Container management platform

Abstract

The invention provides a container management platform. The container management platform comprises a scheduler, a monitoring alarm system, a log processing system and a publishing system, wherein the scheduler is an application scheduling framework written based on a Mesos Restful API and is used for life cycle management of container applications; the monitoring alarm system is used for monitoring performance indicators of containers and health states of the applications; the log processing system is used for log retrieval and log statistics; and the publishing system is used for realizing publishing and rollback of the applications. Through the container management platform, the position of a service can be quickly found, multi-tenant management under cluster resources can be efficiently utilized, the monitoring alarm system can be flexibly set, and log checking under an associated background can be provided.

Description

technical field [0001] The invention relates to the computer field, in particular to a container management platform. Background technique [0002] In a platform as a service (PaaS) platform, it may be necessary to manage a huge number of containers, and these containers are irregularly distributed in different virtual machines, and will also be migrated to other virtual machines according to external commands or changes in internal resources. On-board, while the realization of functions such as application orchestration and service discovery requires the location of a service to be quickly located. At the same time, a container management platform needs to maintain efficient use of cluster resources, and the monitoring system needs to be able to effectively and flexibly monitor container performance indicators and application health status, and provide more comprehensive log information when problems occur for easy viewing of the analysis. [0003] The current container m...

AI Technical Summary

Problems solved by technology

[0003] The current container management platform has the disadvantage of slow positioning speed when the number of containers increases and in the case of large-scale clusters

And when it comes to multi-tenant management, the current container management platform does not have a high utilization rate of cluster resources. At the same time, the traditional monitoring and alarm system is not flexible enough and requires more complex configurations. Complex configurations lead to the risk of low reliability. Moreover, traditional log retrieval does not directly provide log information in the context of correlation, making log analysis inconvenient.

Images