Add 401 to the list of API codes that our k8s client retries on

Summary:
401 would typically not be a retryable error, but a user reported hitting it when they scaled up their cluster, and aws/containers-roadmap#1810 seems to suggest retrying as a workarounds. The downside of retrying on a 401 seems fairly low as well. Open to push-back on this though.

Test Plan: Existing BK test coverage of the k8s client

python_modules/libraries/dagster-k8s/dagster_k8s/client.py

@@ -78,6 +78,8 @@ class DagsterK8sPipelineStatusException(Exception):
     503,  # Service unavailable
     504,  # Gateway timeout
     500,  # Internal server error
+    # typically not transient, but some k8s clusters raise it transiently: https://github.com/aws/containers-roadmap/issues/1810
+    401,  # Authorization Failure
 ]