We make great use of the %D log format string ("The time taken to serve the request, in microseconds"), but what I'd also really love to have is a similar benchmark measuring from when litespeed initiates the external application request and when it gets a complete result.
The problem with %D is it doesn't tell you how much time was spent processing an external application request, especially in situations where the request was queued in some fashion (wasn't immediately handed off to the external app).
This would be really helpful because it would allow us to analyze the access logs and better determine which client networks are causing the most load on the server. I know it's a wall-clock time and not CPU time, but it's still useful.
I guess really what I want is more details exposed to the custom log format, which nail down into specifics about the external processing. That would be awesome
The problem with %D is it doesn't tell you how much time was spent processing an external application request, especially in situations where the request was queued in some fashion (wasn't immediately handed off to the external app).
This would be really helpful because it would allow us to analyze the access logs and better determine which client networks are causing the most load on the server. I know it's a wall-clock time and not CPU time, but it's still useful.
I guess really what I want is more details exposed to the custom log format, which nail down into specifics about the external processing. That would be awesome