| With the continuous development of virtualization technology over the years, cloud computing is being more and more popular. Modern enterprise IT engineers are no longer satisfied just to get virtual machine nodes with computing capabilities through virtualization technology, and gradually turn their attention to a more ideal environment for application development and ultimately achieve the goal of No-Ops. Cloud Foundry is the first excellent open source PaaS product that came in this background.This thesis combined with the current industry mainstream technology direction of distributed systems and laboratory practical projects of various performance indicators of Cloud Foundry, proposed deployment difficulties and imperfect monitoring problems. Based on these problems, the thesis has designed and implemented a rapid automatically deployment and monitoring system for Cloud Foundry.The system implemented automatic deployment tool without the dependent of the underlying infrastructure as a service platform, each component deployed without redundancy by resolving Cloud Foundry release. The system also proposed and implemented non-intrusive way for monitoring including four levels. They are virtual machines, components, container processes and application logs. The system monitor resource usage and process state, ensure high availability and process automatic recovery, enable application log query, warn and alarm resource usage overrun and fault. These four levels ensure the stable of cloud platform, effectively reducing the the difficulty and complexity of operation and maintenance.Finally, the deployment and monitoring system has withstood the test of actual production environments with quick deployment and stable monitoring. |
