Exactly this. Eliminating variable modes and especially additional work or slower paths during a state change like a failure. Another classic is debug logs. Tu

Exactly this. Eliminating variable modes and especially additional work or slower paths during a state change like a failure. Another classic is debug logs. Turning on additional logging when your system is under stress is a great way to increase the impact. Instead always record/emit the telemetry youll likely need and discard later if its unnecessary to retain.