HTTP client

An HTTP client sends one request to a service and shows what came back: the status, the headers and the body. It works the same against a service in any stack.

Finding it

• On the machine: curl is installed on most systems; curl --version confirms it.
• The address and credentials: the README, an example environment file (.env.example), the API's OpenAPI description.
• Requests the project keeps: .http files or a collection, which show the headers and the body a call needs.

Using it

• Turn an HTTP error into a non-zero exit code: curl --fail, or --fail-with-body to keep the error's body.
• Print the status code alongside the body: curl -sS -w '\n%{http_code}\n' <url>; -i adds the headers, -v the request too.
• Bound the wait with --max-time, so a hung service fails the call instead of hanging it.
• Read tokens from the environment, never typed into the command, where they stay in the history and the logs.

Examples: curl, HTTPie, grpcurl (gRPC).