Does omitting the 'last' cursor in Deta Base fetch operations produce latency spikes that mimic connection pool exhaustion?

Determine how Deta Base fetch pagination behavior without the 'last' cursor influences observed request latency and error patterns in deployed micros instances. Deta Space runs on managed Micros instances where the platform governs concurrent request handling, and the SDK abstracts underlying connection lifecycle, making it difficult to distinguish platform‑throttling from client‑side pagination errors. Item size limits and scan‑based query latency grow with base size, which can produce timeouts that resemble connection exhaustion.

The precise goal is to isolate whether intermittent resource‑exhaustion reports stem from cursor‑handling logic in application code or from the platform’s concurrency model, given that traditional connection‑pool diagnostics do not apply to Deta Base’s API. Because the 'last' cursor is a documented pagination feature, its omission or misuse provides a verifiable entry point for this investigation.

Does omitting the 'last' cursor in a fetch loop produce latency spikes that scale with base size? Can observed timeout patterns be differentiated from cursor‑handling failures without runtime instrumentation? Which concurrency thresholds in the Deta Space runtime interact most strongly with unbounded fetch loops?