Mistake: misreading response-time fields and chasing phantom slowness

Mistake: misreading response-time fields and chasing phantom slowness

Apache %D is microseconds, Nginx $request_time is seconds — mix them up and your crawl-latency report is off by a million. Reads:

→ Apache mod_log_config docs — %D = microseconds, %T = whole seconds. Different fields, different scale.
Takeaway: confirm the directive in your config before plotting latency.

→ Nginx log_module docs — $request_time and $upstream_response_time are seconds with millisecond resolution.
Takeaway: label units in your parser so nobody divides by the wrong factor.

★ Pick of the week — a unit-sanity awk guard — Flags impossible values (e.g. 0.000001s pages).
Takeaway: Googlebot throttles on real latency — measure it in the right units.