admin: fixed mapping of service unavailable error code

Previously the service_unavailable error code was mapped to generic
unexpected server error. This lead to a situation in which a lot of
errors were logged when broker was restarted as it has an RPC server
running but without the services registered. Changed the mapping of
`rpc::errc::service_unavailable` to HTTP service unavailable error.

Fixes: redpanda-data#17739

Signed-off-by: Michał Maślanka <michal@redpanda.com>
(cherry picked from commit 1a9f248)

src/v/redpanda/admin/server.cc

@@ -1118,6 +1118,7 @@ ss::future<> admin_server::throw_on_error(
         case rpc::errc::disconnected_endpoint:
         case rpc::errc::exponential_backoff:
         case rpc::errc::shutting_down:
+        case rpc::errc::service_unavailable:
         case rpc::errc::missing_node_rpc_client:
             throw ss::httpd::base_exception(
               fmt::format("Not ready: {}", ec.message()),
@@ -1130,7 +1131,6 @@ ss::future<> admin_server::throw_on_error(
         case rpc::errc::service_error:
         case rpc::errc::method_not_found:
         case rpc::errc::version_not_supported:
-        case rpc::errc::service_unavailable:
         case rpc::errc::unknown:
             throw ss::httpd::server_error_exception(
               fmt::format("Unexpected error: {}", ec.message()));