I trust logs more than I trust the debugger
Stepping through code shows you one run. Good logs show you the pattern.
The debugger is great for one well-understood failure. But most of my hard bugs are about timing, ordering, and the gap between what I think runs and what actually runs. For that, structured logs beat breakpoints every time.
A breakpoint freezes one execution. A log line you can grep across a thousand executions tells you whether the thing you suspect happens always, sometimes, or never — which is usually the actual question.