Monitoring a kontrola aktualizací dat Power BI objektů

The master's thesis deals with the topic of refresh monitoring and data quality assurance in the Microsoft Power BI cloud ecosystem. The aim of the thesis is the description of Business Intelligence architecture, identification of the cascading errors problem, description of the use of Power BI REST API for operational metadata extraction, and verification of the proposed monitoring methods in a practical implementation on a case study. The literature review includes data warehouse principles, data refresh processes, cloud security policies, and the limits of built-in monitoring tools. Furthermore, the thesis focuses on the possibilities of programmatic access to logs with an emphasis on M language development and the use of the On-premises Data Gateway. The practical part of the thesis selects three main types of Power BI data objects (Dataflows, Datamarts, Datasets). It proposes methods for downloading their history using a Custom Data Connector, including the integration of undocumented API endpoints. The proposed solution methods are subsequently verified on a case study for multinational company. In the case study, the connection of isolated metadata into a unified data lineage for cascading error detection is implemented. The refresh status controls are implemented as a relational semantic model using the custom connector and incremental refresh. Subsequently, this relational model is visualized in an interactive report for monitoring and rapid reaction and error correction. The thesis brings valuable insights for organizations managing extensive data portfolios and provides a practical guide for building secure governance using the Power BI REST API without the need to escalate administrative privileges.