[BUG] Communication in the same docker compose about compose

[ndeloof]

For containers to communicate, those must be connected to a common network. Compose creates one by default for this purpose, if you have no reason to set network1/network2 just remote all networks declaration

[wisamidris7]

I initially encountered issues not with SQL Server or myApp, but with gRPC communication between two applications within the same docker-compose setup.

My goal was to configure these applications for both HTTP/1 and HTTP/2 communication. However, I discovered that HTTP/1 and HTTP/2 require HTTPS or TLS for proper functionality.

In my ASP.NET Core project, which utilizes both gRPC and HTTP APIs, everything worked perfectly locally. But upon deploying to Docker, I encountered a problem: HTTP/2 is not supported, and HTTP/1 was required for gRPC calls from myapp1 to myapp2.

To address this, I configured myapp2 to use HTTPS with a domain name and certificate through Kestrel, setting HttpProtocol to Http1AndHttp2. While this configuration worked for requests from Chrome (using HTTP/1) and an empty gRPC console application (using HTTP/2), requests from myapp1 to myapp2 failed. They went through IP address-based communication and resulted in certificate validation errors due to the use of HTTPS.

I attempted to resolve this by changing the URL from https://myapp2:10000 to https://example.com:3912, but this led to timeout issues with gRPC requests. Additionally, I tried isolating each service in separate networks, but this did not resolve the issue.

I am also encountering similar problems when using reverse proxies like Nginx, as it does not seem to support HTTP/1 and HTTP/2 properly.

Given these challenges, I am concerned that this might be a Docker-related issue, especially since the problem persists even when services are in different networks. Any insights or suggestions would be greatly appreciated.