Increase default upper error bound in validate NTP (#477)

The default for the maximum upper clock error boundary in the "Validate
NTP synchronization" playbook was quite low.
Usually in OSISM deployments ceph is one of the most time sensitive
applications. In its default ceph starts to show warnings when it
detects clock drift above 0.05 seconds between monitors [1]. Therefore,
to ensure the quality of NTP time allows to keep the drift between two
systems to less or equal to 0.05 seconds a default of 0.025 seconds will
be used for `max_upper_clock_error_bound`.

[1]
https://docs.ceph.com/en/latest/rados/configuration/mon-config-ref/#confval-mon_clock_drift_allowed

Signed-off-by: Jan Horstmann <[email protected]>

playbooks/generic/validate-ntp.yml

@@ -6,7 +6,15 @@
   vars:
     max_stratum: 9
     max_system_clock_deviation: 0.005
-    max_upper_clock_error_bound: "{{ max_system_clock_deviation + 0.015 }}"
+    # NOTE: Usually in OSISM deployments ceph is one of the most time sensitive
+    # applications. In its default ceph starts to show warnings when it
+    # detects clock drift above 0.05 seconds between monitors [1]. Therefore,
+    # to ensure the quality of NTP time allows to keep the drift between two
+    # systems to less or equal to 0.05 seconds a default of 0.025 seconds will
+    # be used for `max_upper_clock_error_bound`.
+    #
+    # [1]: https://docs.ceph.com/en/latest/rados/configuration/mon-config-ref/#confval-mon_clock_drift_allowed
+    max_upper_clock_error_bound: "{{ max_system_clock_deviation + 0.02 }}"
 
   tasks:
     - name: Get chrony tracking