Check duplicate toleration on pods and PrometheusDuplicateTimestamps alert

ocs_ci/ocs/constants.py

@@ -1394,6 +1394,7 @@
 ALERT_OSDDISKUNAVAILABLE = "CephOSDDiskUnavailable"
 ALERT_PGREPAIRTAKINGTOOLONG = "CephPGRepairTakingTooLong"
 ALERT_PROMETHEUSRULEFAILURES = "PrometheusRuleFailures"
+ALERT_PROMETHEUSDUPLICATETIMESTAMPS = "PrometheusDuplicateTimestamps"
 ALERT_BUCKETREACHINGQUOTASTATE = "NooBaaBucketReachingQuotaState"
 ALERT_BUCKETREACHINGSIZEQUOTASTATE = "NooBaaBucketReachingSizeQuotaState"
 ALERT_BUCKETERRORSTATE = "NooBaaBucketErrorState"

ocs_ci/ocs/exceptions.py

@@ -762,3 +762,7 @@ class ActiveMdsValueNotMatch(Exception):
 
 class DistributionStatusError(Exception):
     pass
+
+
+class DuplicateTolerationException(Exception):
+    pass

ocs_ci/ocs/resources/pod.py

@@ -16,6 +16,7 @@
 from threading import Thread
 import base64
 from semantic_version import Version
+from collections import Counter
 
 from ocs_ci.ocs.ocp import get_images, OCP, verify_images_upgraded
 from ocs_ci.helpers import helpers
@@ -34,6 +35,7 @@
     TimeoutException,
     NoRunningCephToolBoxException,
     TolerationNotFoundException,
+    DuplicateTolerationException,
 )
 
 from ocs_ci.ocs.utils import setup_ceph_toolbox, get_pod_name_by_pattern
@@ -2776,6 +2778,41 @@ def check_toleration_on_subscriptions(toleration_key=constants.TOLERATION_KEY):
         )
 
 
+def check_duplicate_tolerations_on_pods():
+    """
+    Function to check for duplicate tolerations on pods.
+
+    Raises:
+        DuplicateTolerationException: Raised when duplicate tolerations are found on any pod.
+    """
+    pod_objs = get_all_pods(
+        namespace=config.ENV_DATA["cluster_namespace"],
+        selector=[constants.ROOK_CEPH_OSD_PREPARE],
+        exclude_selector=True,
+    )
+
+    pods_with_duplicates = []
+    for pod_obj in pod_objs:
+        resource_name = pod_obj.name
+        tolerations = pod_obj.get().get("spec", {}).get("tolerations", [])
+        toleration_tuples = [tuple(sorted(tol.items())) for tol in tolerations]
+        counts = Counter(toleration_tuples)
+
+        duplicates = [tol for tol, count in counts.items() if count > 1]
+        if duplicates:
+            logger.warning(
+                f"The pod {resource_name} has duplicate tolerations: "
+                f"{[dict(tol) for tol in duplicates]}"
+            )
+            pods_with_duplicates.append(resource_name)
+
+    if pods_with_duplicates:
+        raise DuplicateTolerationException(
+            f"The following pods have duplicate tolerations: "
+            f"{', '.join(pods_with_duplicates)}"
+        )
+
+
 def run_osd_removal_job(osd_ids=None):
     """
     Run the ocs-osd-removal job

tests/functional/z_cluster/nodes/test_toleration.py

@@ -9,17 +9,19 @@
     skipif_managed_service,
     skipif_hci_provider_and_client,
 )
+from ocs_ci.ocs import constants
 from ocs_ci.ocs.resources.pod import (
     get_all_pods,
     check_toleration_on_pods,
     wait_for_pods_to_be_running,
+    check_duplicate_tolerations_on_pods,
 )
 from ocs_ci.ocs.node import (
     get_ocs_nodes,
     taint_nodes,
     untaint_nodes,
 )
-
+from ocs_ci.utility.prometheus import PrometheusAPI
 
 logger = logging.getLogger(__name__)
 
@@ -47,7 +49,7 @@ def finalizer():
 
         request.addfinalizer(finalizer)
 
-    def test_toleration(self):
+    def test_toleration(self, threading_lock):
         """
         1. Check if nodes are tainted
         2. Taint ocs nodes if not tainted
@@ -64,6 +66,17 @@ def test_toleration(self):
         # Check tolerations on pods under openshift-storage
         check_toleration_on_pods()
 
+        # Check duplicate toleration on pods and PrometheusDuplicateTimestamps alert(DFBUGS-1654)
+        check_duplicate_tolerations_on_pods()
+        prometheus = PrometheusAPI(threading_lock=threading_lock)
+        alerts_response = prometheus.get(
+            "alerts", payload={"silenced": False, "inhibited": False}
+        )
+        alerts = alerts_response.json()["data"]["alerts"]
+        assert constants.ALERT_PROMETHEUSDUPLICATETIMESTAMPS not in [
+            alert["labels"]["alertname"] for alert in alerts
+        ]
+
         # Respin all pods and check it if is still running
         pod_list = get_all_pods(namespace=config.ENV_DATA["cluster_namespace"])
         for pod in pod_list: