Custom ranking attribute list without documented safe maximum

A repeatable development environment requires index settings for custom ranking to be versioned and reproduced consistently across dev and staging.

Custom ranking allows an unlimited number of attributes in the configuration. Public documentation does not specify a documented constraint or performance guideline for a safe maximum number of attributes.

Without a defined bound, it is unclear how attribute count influences query latency and relevance stability when the same settings are applied in different environments with different data volumes.

Questions:

• Is there a documented maximum number of custom ranking attributes supported by Algolia?
• What performance guidance exists for query latency as the custom ranking list grows?
• How should a repeatable dev setup cap or validate the attribute list to match production behavior?