Multi-Level Monitoring Visualization

Description

Integriertes / Multi-Level Monitoring  and Visueller Leitstand  are part of the project TUM Living Lab Connected Mobility (TUM LLCM).

The project is located in the field of IT operations analytics (ITOA) which represents an approach or method to retrieve, analyze, and report massive amounts of monitoring data for IT operations. This research focuses in particular on providing proof of concepts and best practices for the implementation of an monitoring and analytic architecture which provides

• historical and real-time information about the usage (load profiling) and quality (workload, availability, response time, throughput, error frequency) of microservice based software systems on different abstraction levels in a consistent manner. The abstraction levels covers the hardware on which the services are deployed, functional (business or core services) and infrastructural (load balancing, proxy, configuration, service discovery) services as well as the user behavior in according to how the user groups interact with the system.
• The correlation or connection between the abstraction levels in order to discover the dependency model of service oriented architectures. This covers in particular how the services communicate (synchron or asynchron) with each other.
• Answers to analytical questions on how to perform efficiently web log mining for predicting user paths and clicks, service anomaly detection and failure impact analysis.

The visual control panel will provide information, which supports the operation and management of microservice based architectures.  “Business Continuity” is one of the critical success factors for the service operator. In order to be able to ensure this, it is required to monitor the availability of the provided services continuously. Because of the diversity of the provided services and the variety of potential control panel user groups, the subproject requires flexible, expendable, target group-specific and configurable data views. Furthermore, the data on technical availability has to be integrated with the transactional business data in order to be able to draw conclusions of the users’ behavior in terms of business performance. Currently, there do not exist consistent modelling approaches for the presentation and integration of such monitoring data.

Research Questions

• What is an appropriate architecture for monitoring and managing cross-layer dependencies in microservice environments?
• Which monitoring tools are required and which technical prerequisites do they have to fullfill?
• How can architecture discovery concepts be extended in order to reconstruct cross-layer dependencies?
• How to analyze and predict the user click path through the monitored application?
• How to uncover failure impacts on other services and the user behavior based on monitoring data and the dependency knowledge?

Objectives

• Best Practices for an Open Mobility Services Platform Monitoring Solution (Microservices Architecture)
• Tightly coupled feedback loop between platform development and platform operation
• Correlation of monitoring data from multiple abstraction layers
• IT operation analytics alongside of customer journeys
• Impact analysis of service failures
  

Use Cases - Work in progress

Prototypes

Publications