{
order trojan before file_server
servers :4443 {
listener_wrappers {
trojan
}
}
log {
output file /var/log/caddy/default.log {
roll_size 10MiB
}
}
trojan {
caddy
no_proxy
users user tro_password
}
}
tro.arm-br.example.com {
reverse_proxy https://tro-arm-br.example.com:4443
}
naive.arm-br.example.com {
reverse_proxy https://arm-br.example.com:443
}
plex.arm-br.example.com {
reverse_proxy arm-br.example.com:32400
}
qbit.arm-br.example.com {
reverse_proxy arm-br.example.com:28080
}
cd2.arm-br.example.com {
reverse_proxy arm-br.example.com:19798
}
status.arm-br.example.com {
reverse_proxy arm-br.example.com:10182
}
:443, arm-br.example.com {
tls [email protected]
route {
forward_proxy {
basic_auth user naive_password
hide_ip
hide_via
probe_resistance
}
file_server {
root /usr/share/caddy
}
}
}
:4443, tro-arm-br.example.com {
tls [email protected]
route {
trojan {
connect_method
websocket
}
file_server {
root /usr/share/caddy
}
}
}
:6443, arm-brv6.example.com {
tls [email protected]
route {
forward_proxy {
basic_auth user naive_password
hide_ip
hide_via
probe_resistance
}
file_server {
root /usr/share/caddy
}
}
}
caddy 一共使用了两个插件,一个是 caddy-trojan ,另一个是 naive-proxy 作者修改的 forwardproxy 插件,几个 vps ,凡是按照上面的配置文件部署了 caddy 的,内存占用都是起码 500-600M ,我 24g 的甲骨文 arm 机器,caddy 内存占用甚至达到了 5-6g ,问了下 chatgpt ,也没给出什么实质性的建议;重启 caddy 之后,内存占用倒是降下来了,但是似乎随着时间增长,占用会逐渐升高,请问要如何排查问题呢?