We have a site of about 3,000 nodes we are indexing to Algolia.
Only a single one of those results is over the 10kb limit -- all others average 1.58kb in size.
When we clear the index and attempt to re-index all of our content, the process halts as soon as it encounters this 20kb+ node. The only error message we see is that object at position X was over the 10kb limit and could not be indexed.
What we DON'T get is any messaging/errors that indicate this failure caused all remaining items to be rejected.
The expected behavior is that when an over-limit or otherwise problematic entity throws an error, that particular entity is logged and skipped, and the indexing process continues on normally with the remaining items.
Comments
Comment #2
julienbourdeau commentedHi,
I understand it would make sense to not fail and display an error message at the end. I guess it "force" the user to act on it, where if most content was indexed, it could be left out.
To index the document, you can either troncate the content (if you have very long description for instance) or you can split the entry into multiple records. I haven't used the drupal library so I'm not sure how to do that in details.
If you split the entry, you can make sure only one result shows up for the page: https://www.algolia.com/doc/tutorials/full-text-search/distinct/
Best regards,
Julien
Comment #3
nikunjkotechaPlease check and revert back if you still face issues with latest version.
Comment #4
maskedjellybeanFor anyone interested in this issue, I created a search API processor which splits field values into multiple Algolia index records based on a character limit. It could use review and improvement: https://www.drupal.org/project/search_api_algolia/issues/3256840
Comment #5
nikunjkotechaClosing in favour of https://www.drupal.org/project/search_api_algolia/issues/3256840