Definition of a Microservices-based Management and Monitoring System for Oracle Cloud

Oracle Cloud Infrastructure (OCI) is a highly performing IaaS (Infrastructure as a service) solution, however unlike other cloud operators (e.g., Aws, Azure, Google), the usage of its portal with its graphical interface is not so immediate and intuitive for an end user. Indeed, features such as resource lifecycle monitoring, as well as cost monitoring, are dispersive and are mostly designed for an expert user. The cloud itself is also partitioned internally; in fact, each company that signs in is provided with an isolated partition called "tenancy" where it is possible to create, organize and manage cloud resources securely. Furthermore, within each tenancy it is also possible to divide the resources featured into logical groups defined as "compartments", each one protected by different policies defined by the user. Starting from the root compartment, it is therefore easy to recognize a tree structure which rests on several levels. Whenever a user intends to carry out any search on a compartment to check the status of the resources allocated on it, the same will be always limited to returning elements belonging to the first level of the structure, without searching in depth in the various sub-levels, hence it will not even be possible have a complete view of all the existing resources in the tenancy or in a portion of it at a given time. The thesis aims to create an application, leveraging a microservice-based architecture, to monitor the entire cloud, oriented to the end user, which can immediately and promptly provide information on the life cycle, as well as on the costs, of individual resources and the possibility of managing their status in real time through a convenient graphical interface. The main goals are summarised below: •Design and development of a simpler and more intuitive GUI (Graphical User Interface) capable of requesting, ordering and presenting the information that the end user needs. The interface will, other than providing a preliminary authentication stage, have to recall a list of REST API developed at back-end to access the differentiated services provided by the architecture and aggregate data that currently are unrelated, such as cost and resource utilization level. • Maintaining the system independently scalable in the face of an unpredictable number of requests from end users, by deploying it on Oracle Cloud Infrastructure Container Engine for Kubernetes (OKE). An application which disposes a number of microservices needs to be highly scalable, having to create many instances for each service and to balance these services on many hosts. All of this involves a huge level of deployment complexity for IT operations and management, as a consequence it's natural to think about using a microservice-oriented infrastructure as well as an orchestrator like Kubernetes. • Searching for the right trade-off between the amount of resources used for system deploying (e.g., CPU, RAM, etc.) and quality of service (QoS) experienced by the end user. Parameters such as latency and number of errors must be kept to a minimum, in this perspective, where it will be possible and convenient, caching strategies will be put in place to store all the information that does not need to be shown updated in real time and which would imply a considerable decrease in response times of individual calls and consequently in total latency.