Retries on connect timeout

Currently connect timeouts are not retried for POST/PATCH requests which are not considered idempotent. I agree that on connection reset we can't tell if the server has already started processing the request or not, but I wonder if there is any reason to not allow retries up to `max-retries` on connection timeout errors (no response to SYN packet) even for POST/PATCH

Trace the HTTP client retry policy and the handling of connection timeout errors alongside the POST/PATCH idempotency rules. Confirm the intended safety trade-off with maintainers, then verify that connection timeouts can retry up to max-retries for these methods while other non-idempotent failures retain their existing behavior.

Written by the indexing model from the issue text.