Consider increasing the default timeout for _bulk to 330s

The Elastic cloud proxy has a 320s timeout. Our current default network timeout to _bulk is 90s. This means there are long running requests that Elasticsearch could plausibly complete in >90s that we give up on and then retry unnecessarily.

At the same time, a longer default timeout will possibly be incompatible with non-Elastic network middleware like on premise proxies or firewalls with shorter connection timeouts than this.

Start by locating the Go code that sets the default network timeout for _bulk requests and the tests covering that default. Review how the timeout interacts with Elasticsearch cloud proxies and non-Elastic middleware, then decide whether the default should be 330s or remain unchanged. Done means the selected default and its compatibility behavior are covered by tests.

Written by the indexing model from the issue text.