Reduce the bulk indexing request size for 429 errors too

[mpfz0r]

An elastic cluster might not have enough memory available to accept a full bulk index request. It can reject it with a HTTP/1.1 429 Too Many Requests exception.

We could handle this the same way we do with Request entity too large errors. See https://github.com/Graylog2/graylog2-server/pull/7071