Optimal chunk size for bulk & scan

We are using Elasticsearch python lib. We are doing a lot of bulk and scan operations. I wonder if there is any recommendation about the optimal bulk size for write/update operations and chuck size for scan read operations.

We suggest aiming for 5MB for _bulk. Work backwards to see what that means for doc count, then try adjusting up and down till you get the best performance.

just want to make sure that the 5MB suggestion is relevant for version 1.4.7 as well :slightly_smiling_face:
Can you approve please?
And what about scan?
What is the simple way to see document size?

The only solid recommendation I can make for 1.4 is to upgrade, it was EOL nearly 2 years ago.

I am not sure if there is comparable suggestions for scroll, but you should definitely run some automated testing.

1 Like

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.