[feature request] max_retry support for parallel_bulk

[TomonoriSoejima]

parallel_bulk does not have max_retries parameter like other methods (streaming_bulk) does.

Please kindly consider adding it.

[mazzma12]

would be nice, but i think it would need a bit of refactor. Currently the call to max_retry is processed in streaming_bulk(), but it should be moved to the inner function _process_bulk_chunk() to do so

[sethmlarson]

You can now configured your client to have a max_retries value via .options(max_retries=...) and pass that client to parallel_bulk. Going to close this one.