fix: add timeout to thinking/loop-QA schedulers (#530)

Co-authored-by: Kimi Agent <kimi@timmy.local>
Co-committed-by: Kimi Agent <kimi@timmy.local>

src/config.py

@@ -254,6 +254,7 @@ class Settings(BaseSettings):
     # When enabled, the agent starts an internal thought loop on server start.
     thinking_enabled: bool = True
     thinking_interval_seconds: int = 300  # 5 minutes between thoughts
+    thinking_timeout_seconds: int = 120  # max wall-clock time per thinking cycle
     thinking_distill_every: int = 10  # distill facts from thoughts every Nth thought
     thinking_issue_every: int = 20  # file Gitea issues from thoughts every Nth thought
     thinking_memory_check_every: int = 50  # check memory status every Nth thought

src/dashboard/app.py

@@ -155,7 +155,17 @@ async def _thinking_scheduler() -> None:
     while True:
         try:
             if settings.thinking_enabled:
-                await thinking_engine.think_once()
+                await asyncio.wait_for(
+                    thinking_engine.think_once(),
+                    timeout=settings.thinking_timeout_seconds,
+                )
+        except TimeoutError:
+            logger.warning(
+                "Thinking cycle timed out after %ds — Ollama may be unresponsive",
+                settings.thinking_timeout_seconds,
+            )
+        except asyncio.CancelledError:
+            raise
         except Exception as exc:
             logger.error("Thinking scheduler error: %s", exc)
 
@@ -175,7 +185,10 @@ async def _loop_qa_scheduler() -> None:
     while True:
         try:
             if settings.loop_qa_enabled:
-                result = await loop_qa_orchestrator.run_next_test()
+                result = await asyncio.wait_for(
+                    loop_qa_orchestrator.run_next_test(),
+                    timeout=settings.thinking_timeout_seconds,
+                )
                 if result:
                     status = "PASS" if result["success"] else "FAIL"
                     logger.info(
@@ -184,6 +197,13 @@ async def _loop_qa_scheduler() -> None:
                         status,
                         result.get("details", "")[:80],
                     )
+        except TimeoutError:
+            logger.warning(
+                "Loop QA test timed out after %ds",
+                settings.thinking_timeout_seconds,
+            )
+        except asyncio.CancelledError:
+            raise
         except Exception as exc:
             logger.error("Loop QA scheduler error: %s", exc)