Optimizing Dynamic Laravel Sitemap Generation: Encountering Memory Exceeded Errors with Extensive Databases

We're currently engaged in building dynamic sitemaps for a Laravel application that features a rapidly expanding database, now encompassing millions of records across various models that absolutely require indexing for optimal search engine visibility. Our critical challenge lies in persistent memory exhaustion and script timeouts during the sitemap generation process itself, particularly when our system attempts to fetch and process the extensive volume of URLs directly from the database. While our existing methodology proved adequate for smaller datasets, it's definitively not scaling with the current data growth, severely impacting our Laravel SEO performance.

We've already implemented standard workarounds such as increasing PHP memory limits and employing basic chunking techniques utilizing Laravel's cursor() method; however, at our current operational scale, these solutions are proving to be merely temporary fixes rather than sustainable long-term strategies. We are now in urgent need of a more robust and truly scalable solution. Therefore, I'm reaching out to the community for expert advice on advanced strategies for achieving highly efficient and truly scalable dynamic Laravel sitemap generation, especially when dealing with such exceptionally large datasets. What are the definitive best practices and architectural patterns for handling these scenarios without consistently crashing the server or exhausting resources? I'm eagerly awaiting any insights from those who've tackled similar complex challenges.