Database management systems such as Db2 for z/OS have numerous measuring instruments and variables at their disposal in order to be able to provide comprehensive information on a regular basis about their status, the data stored in them and the queries processed. Based on this information, monitoring tools make it easy to view and evaluate them. In the context of SQL performance tuning, for example, the CA Detector can be mentioned here. It makes it possible to view the performance of SQL statements (with regard to execution times, waiting times, etc.) in a predefined way and to compare them with historical data. Its visualization capabilities are unfortunately severely limited.
On the basis of the performance data pre-processed by the Detector, own evaluations can be flexibly created in a suitable environment via SQL and visualized with suitable tools (e.g. Excel). However, there is a multiple media break (Detector in the terminal emulation, SQL execution in a separate environment and visualization e.g. with Excel) with various manual steps. Jupyter Notebook could considerably simplify this task by combining text, executable code and dynamic visualizations and enable flexible evaluations “from a single source”.
The aim of the work is to get to know Jupyter Notebook and its possibilities in order to evaluate the feasibility of the above thesis (flexible reports using Jupyter). If the suitability is determined, selected example evaluations are to be realized in Jupyter notebooks and a compact guideline is to be developed on how to proceed with the creation of such evaluations.