Optimize components for large datasets

**Description**
Currently api lacks proper retry support.

**Proposed solution**
Add configurable retry with sensible defaults.

**Priority**: Low/Medium

The title mentions optimizing components for large datasets, while the description asks for configurable API retries. Start by locating the API request entry point and checking how failures are currently handled. Done should mean configurable retry behavior with sensible defaults, backed by tests for retry and non-retry cases.

Written by the indexing model from the issue text.