Caddy has been sending requests

[BOBOKaa]

1. Version 2.4.6

2. How I run Caddy:

a. System environment:

ubuntu 18.04
no docker

b. Command

systemctl start caddy.service

c. Caddyfile

mydomainname {
   reverse_proxy 127.0.0.1:19988

   log {
    output file /var/log/caddy/access.log
   }
}

3. The problem I’m having:

According to the logs, it can be observed that there are hundreds of requests in 1 minute, and I don’t know why

4. Error messages

status caddy

server’slog

{"level":"error","ts":1643879594267802,"logger":"http.log.error.log0","msg":"EOF","request":{"remote_addr":"ip:port","proto":"HTTP / 1.1","method":"GE T","ho st":"x xx","u ri": " ","he a  dedrs":{"Connectdion":["Updgrade"],"Pragdma":["no-cache"],"Cache-Control":["no-cache"],"Upgrade":["websodcket"],"Sec-W ebsdocket-Version"::["13"],"Sec-Wdebsocket-Key":["sJ="],"Sec-Websocket-Extensions":["; client_max_window_bits"],"User-Agent":["*"],"Origin":["url"],"Accept-Encoding":["gzip, deflate, br"],"Accept-Language":["en-US,en;q=0.9,el;q=0.8,fr;q=0.7"]},"tls":{"resumed":false,"version":772,"cipher_suite":4865,"proto":"","proto_mutual":true,"server_name":"uri"}},"duration":0.000359377,"status":502,"err_id":"vf32r9","err_trace":"reverseproxy.statusError (reverseproxy.go:886)"}

[francislavoie]

Looks like your upstream app ran out of available connections. You’ll probably need to make some tweaks to allow more connections. I don’t think this is a problem with Caddy.

You’ll need to dig deeper into your upstream app, where the problem seems to lie.

[BOBOKaa]

Thank you very much,I have another question as to why caddy is reporting the following error, I have a similarly configured environment, but it works properly

{"level":"error","ts":1643879594267802,"logger":"http.log.error.log0","msg":"EOF","request":{"remote_addr":"ip:port","proto":"HTTP / 1.1","method":"GE T","ho st":"x xx","u ri": " ","he a  dedrs":{"Connectdion":["Updgrade"],"Pragdma":["no-cache"],"Cache-Control":["no-cache"],"Upgrade":["websodcket"],"Sec-W ebsdocket-Version"::["13"],"Sec-Wdebsocket-Key":["sJ="],"Sec-Websocket-Extensions":["; client_max_window_bits"],"User-Agent":["*"],"Origin":["url"],"Accept-Encoding":["gzip, deflate, br"],"Accept-Language":["en-US,en;q=0.9,el;q=0.8,fr;q=0.7"]},"tls":{"resumed":false,"version":772,"cipher_suite":4865,"proto":"","proto_mutual":true,"server_name":"uri"}},"duration":0.000359377,"status":502,"err_id":"vf32r9","err_trace":"reverseproxy.statusError (reverseproxy.go:886)"}

[francislavoie]

EOF just means “Caddy wasn’t able to read a response from upstream”. This means the upstream closed the connection before Caddy was able to receive a response and it ran out of bytes to read when it was expecting more.

[BOBOKaa]

thank you,
It didn’t seem like caddy’s problem, and I dug into it

[BOBOKaa]

What this sentence says？

dial tcp: lookup 3s on 127.0.0.1:53: read udp 127.0.0.1:54508->127.0.0.1:53: read: connection refused
dial tcp: lookup ip_hash on 127.0.0.1:53: read udp 127.0.0.1:40466->127.0.0.1:53: read: connection refused
dial tcp: lookup 400: no such host

[francislavoie]

That means Caddy wasn’t able to resolve DNS for domains named 3s, ip_hash or 400.

I think you made a configuration mistake, but you didn’t share your config so I can’t really suggest anything. Make sure you used the right syntax for reverse_proxy.

[BOBOKaa]

mydomainname {
   reverse_proxy 127.0.0.1:19988
{
       lb_policy ip_hash
       health_path /
       health_status 400
       health_interval 5s
       health_timeout 3s
   }

   log {
    output file /var/log/caddy/access.log
   }
}

[francislavoie]

Make sure your { is on the same line as the directive it pairs with. Syntax is important.

Also, using lb_policy and active health checks won’t give you much benefit if you only have a single upstream. The benefits of those are when you have multiple upstreams that you need to load balance between.