This blog post discusses implementing a timer for multi-statement operations in YugabyteDB or PostgreSQL, providing an alternative to the native "\timing on" feature. The author highlights the limitations of "\timing on", such as not supporting end-to-end timing and not storing results in a spool file. To address this, they propose using SQL to implement a stopwatch that reads the wall clock at the start and finish of operations, allowing for end-to-end timing and recording the measured time in a spool file. The author provides several schemes for implementing the memo, including using an ordinary table, a user-defined run-time parameter, a ysqlsh variable, or an ordinary PL/pgSQL variable within a procedure. They compare the noise levels of these approaches and recommend choosing the best method based on the specific use case. Ultimately, they suggest installing the procedure start_stopwatch() and function stopwatch_reading() centrally in any database to make ad hoc end-to-end timing easy for users.