Skip to content

Secure, Fast, and Extensible Sandbox runtime for AI agents.

docs: explain intentional polling error suppression

贾岛7fe032a369aa
diff --git a/examples/aio-sandbox/main.py b/examples/aio-sandbox/main.py
index 39b016b1e..964decdd3 100644
--- a/examples/aio-sandbox/main.py
+++ b/examples/aio-sandbox/main.py
@@ -51,6 +51,7 @@ def check_aio_process(sbx: SandboxSync) -> bool:
print(f"[check] sandbox ready after {elapsed:.1f}s")
return True
except Exception:
+ # The process may still be starting; retry quietly until polling times out.
pass
time.sleep(0.2)
return False
diff --git a/examples/openclaw/main.py b/examples/openclaw/main.py
index a8e70c208..3e9d2d32d 100644
--- a/examples/openclaw/main.py
+++ b/examples/openclaw/main.py
@@ -55,6 +55,7 @@ def check_openclaw(sbx: SandboxSync, port: int = DEFAULT_PORT) -> bool:
print(f"[check] sandbox ready after {elapsed:.1f}s")
return True
except Exception:
+ # The gateway may still be starting; retry quietly until polling times out.
pass
time.sleep(0.2)
return False
diff --git a/kubernetes/hack/pool-perf.py b/kubernetes/hack/pool-perf.py
index d67864ae2..5b4b2340f 100755
--- a/kubernetes/hack/pool-perf.py
+++ b/kubernetes/hack/pool-perf.py
@@ -129,6 +129,7 @@ class PoolPerformanceTester:
self.results[name]["allocated_time"] = time.time() - start_polling
break
except Exception:
+ # Keep polling through API/read failures until allocation succeeds or times out.
pass
await asyncio.sleep(self.poll_interval)