On Fri, May 27, 2022 at 7:34 PM Eric Covener <cove...@gmail.com> wrote: > > These types of investigations can be really painful. Is there any > harm in allowing the server to act like "KeepAliveTimeout 5" is e.g. > "KeepAliveTimeout 5200ms". > > If this fudge buffer existed as an addl directive (rather than a trick > documented in KeepAliveTimeout) , would it be reasonable as a non-zero > default to discourage this race?
Sounds reasonable to me. Possibly mpm_event when busy should not kill connections in keep-alive state for less than the timeout bias (200ms in your example) too? Regards; Yann.