19/03/2024
👉 Web Request Flow for Git Fetch Operation:
Git operations over HTTP employ the stateless “smart” protocol while processing the fetch request from the user.
However, the responsibility for handling these operations is distributed across several GitLab components.
👉 Here's a simplified representation of the Git fetch operation sequence, relative to a /namespace/project.git URL, with NGINX and other load balancers facilitating the process without altering the requests:
Initial Request Handling:
🚀 The client sends a fetch request to the server's Git repository.
🔄 NGINX forwards the request.
Authentication Check:
🔒 Request undergoes authentication validation.
Fetch-Pack Operation:
📦 Client requests available references and object transfers.
Info-refs Request:
📝 Client seeks metadata via an info-refs request.
SmartHTTPService Handling:
🛠️ SmartHTTPService manages request processing and communication.
Git Upload-Pack Process:
📤 Server invokes git upload-pack to provide metadata and references.
Response Handling:
📨 SmartHTTPService delivers the response to the client via NGINX.
Completion and Status Code:
✅ The operation concludes with a status code confirmation (e.g., 200 OK).
