Routing problem / insufficient debug logs


I'm just in first contact with traefik, using the docker image of version 2.6.3 and following the example with traefik/whoami in docker-compose.

Works as expected with http, but not with https:

http -p HBhb --verify=no http://whoami.localhost

gives the expected answer from the whoami service, but

http -p HBhb --verify=no https://whoami.localhost

which is identical exapt for using TLS results in a HTTP/1.1 404 Not Found

The traefik daeshboard on port 8080 confirms, that the router whoami@docker is configured for both web and websecure entrypoints. with the same rule Host(whoami.localhost). I therefore would expect it to behave identical on both ports 80 http and 443 https, and in both cases the client sends the same request header:

ET / HTTP/1.1
Accept: /
Accept-Encoding: gzip, deflate
Connection: keep-alive
Host: whoami.localhost
User-Agent: HTTPie/1.0.3

Since it runs over the same router, I would expect it to behave similar.

In order to understand it, I set the logging to debug, and it outputs a lot of debug info, but on the request it just says

server_1 | time="2022-03-31T20:29:05Z" level=debug msg="Serving default certificate for request: "whoami.localhost""

nothing else. This is insufficient for debugging. Useless debugging output.

However, the startup debug output says

server_1 | time="2022-03-31T20:35:59Z" level=debug msg="No entryPoint defined for this router, using the default one(s) instead: [web websecure]" routerName=whoami

server_1 | time="2022-03-31T20:36:16Z" level=debug msg="Creating load-balancer" routerName=whoami@docker serviceName=whoami-traefik entryPointName=web

but there is no equivalent for websecure.

So although the whoami service does not specify an entrypoint, traefik claims to connect it to both entrypoints, the dashboard confirms it beeing connected to both entrypoings, there is no log message that a load-balancer router entry for the websecure (https) port has ever been created.

Why not?

And why do the debug log output and dashboard give contradicting information?


Found the problem. Did not configure correctly to ignore SNI, which would not work out of the box.

However, debug log output did not help in finding that problem.