Make the seeder command more resilience to connection errors

[josecelano]

Relates to:

• New console command to upload random torrents to the Index: seeder #453
• Review missing timeouts #467

The seeder command panics when connection is refused by the server.

2024-02-08T12:00:42.181187698+00:00 [seeder][INFO] Uploading torrent #490 ...
2024-02-08T12:00:42.181278716+00:00 [seeder][INFO] Uploading torrent with uuid: 06003a9e-d76e-4ca6-8b45-f05d22a817a3 ...
thread 'main' panicked at /home/josecelano/Documents/git/commiter/me/github/torrust/torrust-index/src/web/api/client/v1/client.rs:181:18:
called `Result::unwrap()` on an `Err` value: reqwest::Error { kind: Request, url: Url { scheme: "http", cannot_be_a_base: false, username: "", password: None, host: Some(Domain("index.torrust-demo.com")), port: None, path: "/api/v1/category", query: None, fragment: None }, source: hyper::Error(Connect, ConnectError("tcp connect error", Os { code: 111, kind: ConnectionRefused, message: "Connection refused" })) }
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

We should catch that error and retry later. This is related to #467. The API Client should return an error instead of panicking.

[josecelano]

After merging #482 now we can catch timeout errors from the API client and retry (N times) with an interval of I seconds between attempts.