fix: Add exponential backoff for shared RQ propagation delay in integration tests

In shared request queue mode, the Apify API has an eventual consistency
propagation delay before newly added or reclaimed requests become visible
in fetch_next_request(). Many integration tests were missing the
call_with_exp_backoff() retry pattern, causing sporadic failures.

Also added rq_access_mode parameter to call_with_exp_backoff() so callers
can pass it directly instead of wrapping every call in an if/else block.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Author: vdusek

tests/integration/_utils.py

@@ -13,13 +13,23 @@
 T = TypeVar('T')
 
 
-async def call_with_exp_backoff(fn: Callable[[], Awaitable[T]], *, max_retries: int = 3) -> T | None:
+async def call_with_exp_backoff(
+    fn: Callable[[], Awaitable[T]],
+    *,
+    rq_access_mode: str | None = None,
+    max_retries: int = 3,
+) -> T | None:
     """Call an async callable with exponential backoff retries until it returns a truthy value.
 
     In shared request queue mode, there is a propagation delay before newly added, reclaimed, or handled requests
     become visible in the API (see https://github.com/apify/apify-sdk-python/issues/808). This helper retries with
     exponential backoff to handle that delay in integration tests.
+
+    When `rq_access_mode` is provided and is not `'shared'`, the function is called once without retries.
     """
+    if rq_access_mode is not None and rq_access_mode != 'shared':
+        return await fn()
+
     result = None
 
     for attempt in range(max_retries):