I recognize "Document tree depth limit exceeded" from Nokogiri (Ruby's HTML parser) actually, might not be from Typesense itself. This could happen if Hudu is tying to process heavily nested HTML content (eg: copy-pasted Word docs or web pages) in order to index it. If that's the case, checking your Hudu app container logs during a reindex should point to the offending document. The
Nokogiri HTML5 docs mention there are options to remove the limit too. I'm new to Hudu though myself, so not sure how you'd feed it through to Nokogiri 😅