Prometheus: Ignore failed SSO jobs (see single-sign-on#26)

test/pytest/test_prometheus.py

@@ -49,12 +49,15 @@ def ignore_alert(alert):
     Returns true if an alert should be ignored. This is when the "severity"
     equals "none", or in some application specific cases.
     """
+
     if alert["labels"]["severity"] == "none":
         return True
+
     # Ignore `KubeAPILatencyHigh` fom high load during installation
     # phase
     if alert["labels"]["alertname"] == "KubeAPILatencyHigh":
         return True
+
     # Filter out failing Nextcloud installation jobs since a lot of
     # them fail until they succeed during installation
     if "nextcloud" in alert["labels"]["pod"]:
@@ -63,6 +66,7 @@ def ignore_alert(alert):
                 "KubeJobCompletion",
                 "KubePodNotReady"]:
             return True
+
     # Filter out when RocketChat pods take very long to start
     if "rocketchat" in alert["labels"]["pod"]:
         if alert["labels"]["alertname"] in [
@@ -70,6 +74,16 @@ def ignore_alert(alert):
                 "KubeDeploymentReplicasMismatch",
                 "KubeStatefulSetReplicasMismatch"]:
             return True
+
+    # Filter out signgle-sign-on job pods until
+    # https://open.greenhost.net/openappstack/single-sign-on/issues/26 is fixed
+    if "single-sign-on-create-" in alert["labels"]["pod"]:
+        if alert["labels"]["alertname"] in ["KubePodNotReady"]:
+            return True
+    if "single-sign-on-create-" in alert["labels"]["job_name"]:
+        if alert["labels"]["alertname"] in ["KubeJobFailed"]:
+            return True
+
     return False