Red Hat Satellite manifest import and refresh failures on very large environments

Issue

• Satellite manifest import and refresh failures on very large environments where tens of thousands of hosts are using a single subscription.

• The import step will fail when Candlepin encounters read timeouts waiting for one of the update steps because it takes too long to complete.

RestClient::Exceptions::ReadTimeout
Katello::Resources::Candlepin::Owner: Timed out reading data from server (POST /candlepin/owners/{Organization}/imports)

Actions::Candlepin::Owner::Import
Katello::Errors::CandlepinError: Runtime Error This connection has been closed. at org.postgresql.jdbc.PgConnection.checkClosed:767