A flexible script that renders a system metrics table using psutil. Function calls are dynamically defined and executed at runtime via a YAML configuration, allowing high customization and adaptability.
Since I wanted to play around with Python, I built this small project focused on OS-level metrics. The idea was to showcase how injecting values via a config file can make code more flexible and adaptable.
Since psutil metric-gathering functions are generated at runtime, we can dynamically add new ones to be rendered in the table. Of course, this comes with a trade-off: higher complexity and trickier debugging. In a real production setup, you'd typically send these metrics to a collector like Prometheus instead-but hey, this is just for fun and learning!
Here’s an example of the generated table using config.yaml:
And another example using config_system_info.yaml, which also includes static system metrics:
Based on the provided config file, we can define the actual values shown, the table’s style, and which metrics should be visualized! 🚀
While this project serves as a proof of concept, here are some potential improvements:
-
More complex calculations: Calculations of provided metrics. This could be implemented via a dynamic calculator that determines operations defined in the YAML file and executes them.
-
Structured logging: In a production setup, structured logging (JSON format) would improve filtering and flexibility in distributed setups.
-
Unit tests: Currently no tests are defined for the DynamicMetric functions.