archival: remove upload retry limit for segments

In a previous PR[1] we began to rely on the archiver loop to retry, and
moved away from relying on `cloud_io::remote` for retries in two ways:
1. setting an explicit `disallow` retry policy on the retry node passed
   to the remote, and
2. setting the `max_retries` passed to `remote::upload_segment()` to 1.

In practice, we saw that _not_ relying on the remote resulted in an
uptick in the `vectorized_cloud_storage_failed_uploads` metric, which is
monitored and alerted on. In [2] we reverted #1, but didn't notice #2.
This commit reverts #2.

[1] redpanda-data#25951
[2] redpanda-data#26969

(cherry picked from commit 7f409da)

Author: andrwng

src/v/cluster/archival/ntp_archiver_service.cc

@@ -1554,13 +1554,7 @@ ss::future<ntp_archiver_upload_result> ntp_archiver::upload_segment(
     // index in the background.
     auto upload_segment_ready = co_await ss::coroutine::as_future(
       _remote.upload_segment(
-        get_bucket_name(),
-        path,
-        meta.size_bytes,
-        get_stream,
-        rtc,
-        lazy_abort,
-        1));
+        get_bucket_name(), path, meta.size_bytes, get_stream, rtc, lazy_abort));
 
     // As noted above, check whether 'get_stream' was called. If not, close the
     // upload stream.