Summary
Understanding the internal state of a running system can be vital to maintaining a high performance, stable system, but conventional approaches such as logging and error handling only expose so much. This talk will touch on how to instrument Python programs in order to observe the state of the system, measure performance, and identify ongoing problems.