OpenShiftUpdateRiskMightApply: bump pending to 15m from 10m

This is to follow up a recent finding [1].

Further digging shows that the related risks to our e2e tests are `TestAlertFeatureE2ETestOTA1813`, `SyntheticRiskA`.

They are in the pending state. This pull bump to a longer pending
time so that it gives more time window to get e2e finish. It could
avoid disruption from Production alerts.

We could use `max by (namespace, risk, reason) (last_over_time(cluster_version_risk_conditions{job="cluster-version-operator", condition="Applies", risk!~"TestAlertFeatureE2ETest.*"}[5m]) != 0)` to ignore testing alerts,
but it does not look good to have code handling special cases only for testing.

This will recover the health of the TP-enabled jobs in CI.

[1]. #1367 (comment)

[2]. openshift/origin#30929

Author: hongkailiu

install/0000_90_cluster-version-operator_02_prometheusrule_servicemonitor.yaml

@@ -16,11 +16,11 @@ spec:
     rules:
     - alert: OpenShiftUpdateRiskMightApply
       annotations:
-        summary: The cluster might have been exposed to the conditional update risk for 10 minutes.
+        summary: The cluster might have been exposed to the conditional update risk for 15 minutes.
         description: The conditional update risk {{ "{{ $labels.risk }}" }} might apply to the cluster because of {{ "{{ $labels.reason }}" }}, and the cluster update to a version exposed to the risk is not recommended. For more information refer to 'oc adm upgrade'.
         runbook_url: https://github.com/openshift/runbooks/blob/master/alerts/cluster-version-operator/OpenShiftUpdateRiskMightApply.md
       expr: |
         max by (namespace, risk, reason) (last_over_time(cluster_version_risk_conditions{job="cluster-version-operator", condition="Applies"}[5m]) != 0)
-      for: 10m
+      for: 15m
       labels:
         severity: warning