Search: index in small chunks

[stsewd]

The default is 500 documents, our documents have nested objects, that can make the final document huge, and timeout ES.

closes #10911

Ref #10911 (comment)

[humitos]

We are going down from 500 to 25, which seems a pretty big drop. Instead of reducing it in 25x I would start by a factor of 5x and use 100 documents.

[sentry-io]

Suspect Issues

This pull request was deployed and Sentry observed the following issues:

• ‼️ ConnectionTimeout: Connection timed out during request readthedocs.projects.tasks.search.index_build View Issue

_{Did you find this useful? React with a 👍 or 👎}