Although any deadline that may exist on a call is logged in aggregate via metrics, for troubleshooting specific calls which have timeout errors, it would be helpful to also log a deadline field via the Zap logger as well (e.g. zap.Duration("deadline", time.Duration(...)) so that failures caused primarily by short incoming deadlines can be triaged more easily.
jquirke
commented
1 year ago
Although any deadline that may exist on a call is logged in aggregate via metrics, for troubleshooting specific calls which have timeout errors, it would be helpful to also log a deadline field via the Zap logger as well (e.g. zap.Duration("deadline", time.Duration(...)) so that failures caused primarily by short incoming deadlines can be triaged more easily.