WAGERBABE DOCS
All Stories
3-4-api-request-batching-deduplicationDoneEpic 3.4

Story 3.4: API Request Batching & Deduplication

Status: done

Tasks

  • Task 1: Create RequestDeduplicator Class (AC: 1, 2, 6, 9, 10) - ALREADY EXISTS (Story 2-2-1)
    • 1.1 `server/app/core/request_coalescer.py` - 352 lines, complete implementation
    • 1.2 `_pending: Dict[str, PendingRequest]` tracking with asyncio.Future
    • 1.3 `asyncio.Lock` for thread-safe access
    • 1.4 `get_or_fetch(key, fetch_fn)` method with coalescing logic
    • 1.5 Error propagation via `future.set_exception()`
    • 1.6 Timeout via `asyncio.wait_for(future, timeout=max_wait)` (default 5s)
    • 1.7 Cleanup in finally block with `del self._pending[cache_key]`
  • Task 2: Integrate with Cache Layer (AC: 5) - ALREADY INTEGRATED
    • 2.1 `fixtures.py` uses combined cache + coalescing + circuit breaker flow
    • 2.2 Order: SWR cache check → coalescer.get_or_fetch() → circuit breaker → API
    • 2.3 Cache populated before notifying waiters via SWR pattern
    • 2.4 Singleton pattern via `get_request_coalescer()`
  • Task 3: Request Batching (AC: 3, 4) - SKIPPED (Deduplication provides most value)
    • 3.1-3.5 Optic Odds API does not support batch endpoints
    • Request coalescing achieves same goal without API batch support
  • Task 4: Metrics Collection (AC: 7, 8) - IMPLEMENTED
    • 4.1 `CoalescerStats` dataclass: total_requests, coalesced_requests, api_calls_made
    • 4.2 `api_calls_saved` property = coalesced_requests
    • 4.3 N/A - using coalescing, not batching
    • 4.4 In-memory stats (Redis persistence not needed for monitoring)
    • 4.5 Debug logging for coalescing events
  • Task 5: Statistics Endpoint (AC: 11) - CREATED
    • 5.1 Created `/api/v1/shared/deduplication/stats` endpoint
    • 5.2 Returns: coalesced_requests, api_calls_saved, pending_requests, hit_rate_percent
    • 5.3 Also added `/api/v1/shared/deduplication/health` and `/stats/reset`
  • Task 6: Testing & Validation (AC: 1-11) - 22 TESTS PASSING
    • 6.1 Unit tests for RequestCoalescer: test_request_coalescer.py
    • 6.2 Load test: test_thundering_herd_prevention (100 concurrent → 1 API call)
    • 6.3 Integration test: fixtures.py uses cache + coalescer + circuit breaker
    • 6.4 Error propagation: test_error_propagates_to_all_waiters
    • 6.5 Timeout: test_max_wait_timeout
    • 6.6 Thread safety: concurrent modifications tested
    • 6.7 Quota monitoring via stats endpoint

Progress

Tasks6/6
Acceptance Criteria0
Total Tasks6

Navigation

View Epic 3.4All Stories