Optimizing Search
To optimize your Search experience, we strongly recommend that you follow these guidelines:
-
Configure Search to index only attributes that you know will be needed for your implementation.
-
Use one source per site domain being crawled. To extract multiple tags from a single source, apply multiple extractors to that source.
-
Reindex a source to update indexes with changes made to domain-level search settings. Recrawl a source to update content or recreate indexes. Follow Best practices for indexing.
-
Ensure that your API tier aligns with your required ingestion rate.
-
Avoid indexing images. Instead, index metadata or short descriptions. If an image is needed for your implementation, index a link to that image.
-
Avoid concurrent or parallel updates to a source.
-
In a live Search implementation, do not use the PROD domain for maintaining sources that point to staging/testing domains.
-
Do not mark tags as aggregated if they are to be used in textual relevance searches. The API will interpret aggregated tags as an array of objects, and ignore them in its response.
-
Do not use content within a file for Search features other than textual relevance.
-
When you configure a web crawler, do not include more than 40 locales per crawler source.
-
When you configure an API push source, do not include more than 30 locales per API push source.
-
Do not assign duplicate IDs to documents in different entities.
-
Document IDs must match the
^[a-zA-Z\d_-]+$regex pattern, and must not exceed 256 characters.