High latency observed when increasing limit parameter in Facebook Graph API pagination

When retrieving large collections through the Facebook Graph API, developers often increase the limit parameter to reduce the number of round‑trips. However, raising the limit can also increase server‑side processing and network transfer time, potentially creating a bottleneck before any optimization is applied. The exact point where latency begins to grow disproportionately is not documented, and cursor‑based pagination adds uncertainty about token validity during rapid data mutations. Understanding this inflection point is essential for deciding whether to tune page size or adjust request frequency.

At what limit value does the average response time start to increase sharply? How does frequent data mutation impact the reliability of the 'after' cursor for subsequent pages? What is the optimal limit that balances reduced request count with acceptable latency for typical payload sizes?