[agentserver-responses] Keep newest history item IDs when applying limit (#48560)

* Keep newest history item IDs when applying limit

get_history_item_ids in the in-memory and file response stores used
resolved[:limit], which dropped recent turns. Slice from the end so
conversation history keeps the newest N items.
Fixes #48514.

* Keep newest history ids under the limit and fix changelog

* Restore an unreleased changelog section for the history id limit fix.

Keep 2.1.0b1 as previously released and name FileResponseStore in the 2.1.0b2 notes.

* Bump azure-ai-agentserver-responses to 2.1.0b2 so Analyze validates the unreleased changelog.

* Clarify that limit keeps newest item IDs from the resolved chain.

---------

Co-authored-by: Shiva S <shivakishore14@gmail.com>

Author: Hashim1999164

sdk/agentserver/azure-ai-agentserver-responses/CHANGELOG.md

@@ -1,5 +1,11 @@
 # Release History
 
+## 2.1.0b3 (Unreleased)
+
+### Bugs Fixed
+
+- `get_history_item_ids` on `InMemoryResponseProvider` and `FileResponseStore` now keeps the newest item IDs when applying `limit`. (#48514)
+
 ## 2.1.0b2 (2026-08-20)
 
 ### Bugs Fixed

sdk/agentserver/azure-ai-agentserver-responses/azure/ai/agentserver/responses/_version.py

@@ -4,4 +4,4 @@
 # Copyright (c) Microsoft Corporation. All rights reserved.
 # ---------------------------------------------------------
 
-VERSION = "2.1.0b2"
+VERSION = "2.1.0b3"

sdk/agentserver/azure-ai-agentserver-responses/azure/ai/agentserver/responses/store/_file.py

@@ -539,19 +539,21 @@ async def get_history_item_ids(
           responses in that conversation and contributes their
           ``history_item_ids + input_item_ids + output_item_ids``.
         - Both may be set; results are concatenated in the same order.
+        - When over ``limit``, keeps the most recent N item IDs from the
+          resolved chain, preserving chronological order in the returned slice.
 
         Deleted responses are skipped (matches the in-memory provider).
 
         :param previous_response_id: Optional response id to chain history from.
         :type previous_response_id: str | None
         :param conversation_id: Optional conversation id to scope history lookup.
         :type conversation_id: str | None
-        :param limit: Maximum number of history item ids to return.
+        :param limit: Maximum number of item IDs to return (most recent N).
         :type limit: int
         :keyword context: Platform context (accepted but unused —
             matches :class:`InMemoryResponseProvider`).
         :paramtype context: PlatformContext | None
-        :returns: List of history item ids (possibly empty).
+        :returns: List of item IDs from the resolved chain (possibly empty).
         :rtype: list[str]
         """
         del context
@@ -579,7 +581,9 @@ async def get_history_item_ids(
 
             if limit <= 0:
                 return []
-            return resolved[:limit]
+            # Keep the most recent N item IDs from the resolved chain,
+            # preserving chronological order in the returned slice.
+            return resolved[-limit:]
 
     # ------------------------------------------------------------------
     # Internal helpers (must be called with self._lock held)

sdk/agentserver/azure-ai-agentserver-responses/azure/ai/agentserver/responses/store/_memory.py

@@ -286,20 +286,22 @@ async def get_history_item_ids(
         *,
         context: PlatformContext | None = None,
     ) -> list[str]:
-        """Resolve history item IDs from previous response and/or conversation scope.
+        """Resolve item IDs from previous response and/or conversation scope.
 
-        Collects history item IDs from the previous response chain and/or
-        all responses within the given conversation, up to *limit*.
+        Collects history, input, and output item IDs from the previous
+        response chain and/or all responses within the given conversation.
+        When over *limit*, keeps the most recent N item IDs from the
+        resolved chain, preserving chronological order in the returned slice.
 
         :param previous_response_id: Optional response ID to chain history from.
         :type previous_response_id: str | None
         :param conversation_id: Optional conversation ID to scope history lookup.
         :type conversation_id: str | None
-        :param limit: Maximum number of history item IDs to return.
+        :param limit: Maximum number of item IDs to return (most recent N).
         :type limit: int
         :keyword context: Platform context for multi-tenant partitioning.
         :paramtype context: ~azure.ai.agentserver.responses.PlatformContext | None
-        :returns: A list of history item IDs within the given scope.
+        :returns: A list of item IDs from the resolved chain.
         :rtype: list[str]
         """
         async with self._locked():
@@ -325,7 +327,9 @@ async def get_history_item_ids(
 
             if limit <= 0:
                 return []
-            return resolved[:limit]
+            # Keep the most recent N item IDs from the resolved chain,
+            # preserving chronological order in the returned slice.
+            return resolved[-limit:]
 
     async def create_execution(self, execution: ResponseExecution, *, ttl_seconds: int | None = None) -> None:
         """Create a new execution and replay container for ``execution.response_id``.

sdk/agentserver/azure-ai-agentserver-responses/tests/unit/test_file_response_store_parity.py

@@ -327,7 +327,8 @@ async def test_history_respects_limit(tmp_path: Path) -> None:
             history_item_ids=["hist1", "hist2"],
         )
         ids = await provider.get_history_item_ids("r_prev", None, limit=3)
-        assert ids == ["hist1", "hist2", "in1"]
+        # Chronological order is oldest-first; truncation must keep the newest IDs.
+        assert ids == ["out1", "out2", "out3"]
         # Non-positive limit returns empty.
         ids_zero = await provider.get_history_item_ids("r_prev", None, limit=0)
         assert ids_zero == []

sdk/agentserver/azure-ai-agentserver-responses/tests/unit/test_in_memory_provider_crud.py

@@ -374,6 +374,8 @@ def test_history__respects_limit() -> None:
 
     ids = asyncio.run(provider.get_history_item_ids("resp_lim", None, 3))
     assert len(ids) == 3
+    # Chronological order is oldest-first; truncation must keep the newest IDs.
+    assert ids == ["in_lim_7", "in_lim_8", "in_lim_9"]
 
 
 def test_history__zero_limit_returns_empty() -> None: