For the last couple of decades, the demand for reliable software has been growing. In order to measure software application's reliability, dozens of reliability analysis methods have been proposed. The accuracy and applicability of the methods is highly dependent on the availability and quality of the failure data collected from the analysed software.

The goal of this thesis is to provide a comprehensive overview of the prevalent failure data collection and prediction techniques.

The thesis shall define what kind of failure data can be collected from software applications and what kind of information can be derived from the data. It shall identify possible information sources available during software life cycle, from which this data can be obtained. It shall discuss their features, advantages and limitations. Afterwards, the thesis shall describe, compare and classify existing failure data collection and prediction approaches and their respective methods.

The practical part of the thesis will consist of a case study. Selected methods shall be applied on an open source application of the student's choice. If possible, results of different methods shall be compared. If any selected method is not applicable on the software, proper reasoning shall be given and possible modifications or alternatives shall be discussed.