Please enable a larger buffer for the access_log and flush timeout

[tkittich]

Is your feature request related to a problem? Please describe. It seems the default access_log writes to disks every 2-3 seconds as monitored by iotop. This might be too often and could be reduced. Large number of disk writes are particularly not good for SSD and NVME disks.

Describe the solution you'd like Please enable a larger buffer for the access_log and flush timeout, e.g., access_log logfile proxy buffer=1024K flush=5m;. This should reduce the number of unnecessary disk writes.