Recently we have surfered a 2 minutes downtime in one Apache Server 2.4. There is not any significant error in error_log like "Max server Limit reached" and neither in acces_log.
Moreover, there is not a huge amount of client request in access_log, and neither bots requests.
This is the behaviour of server-status.
We can see, how Apache Idle Workers decrease to zero before increasing. Notice that nagios monitoring check this each 5 minutes, so we have to consider that there may be some delay and the data may not be 100% reliable.
Finally, this is our server worker configuration:
<IfModule mpm_prefork_module>
StartServers 13
MinSpareServers 10
MaxSpareServers 150
ServerLimit 1024
MaxClients 1024
MaxRequestWorkers 1024
MaxRequestsPerChild 10000
</IfModule>
Should we consider to increase the "MaxSpareServers" despite this is an unrecommnended practice in Apache documentation?
Does anyone have any clues on some other issue that I am overlooking?
Thanks a lot!