Details
-
Bug
-
Status: Accepted
-
Normal
-
Resolution: Unresolved
-
PUP 6.19.0
-
None
-
None
-
Night's Watch
-
Needs Assessment
-
Needs Assessment
Description
Puppet Version: 6.19
Puppet Server Version: 6.14
OS Name/Version: RHEL 6/7/8
PUP-10371 introduced a new metric, 'startup_time', to improve accounting of total puppet runtime. However if you run puppet with '–splaylimit' the 'startup_time' metric will include the time spent in splay which will make the puppet runtime metric quite useless.
And, since we run puppet agent via cron (with splay) at the same time on all nodes, all puppet reports will have the same timestamp - not the time when the puppet run actually begins but the time when the agents splay starts. We send our reports to Foreman and this behaviour makes it very hard to know when puppet agent was actually run.
Desired Behavior:
'startup_time' metric should not include splay time.
Â
Attachments
Issue Links
- relates to
-
PUP-5823 Puppet agent with splay on windows begets random interval between runs
-
- Accepted
-