In this instance the java processes on the data aggregator never got the updated time from when chrony fixed it.
We knew this by checking the /opt/IMDataAggregator/data/performance-spool directory. This is the incoming polled data.
The files had a current time but looking at the ifstats_rate file it had time stamps that were in the future
Cycling the data collector resolved the issue:
- service dcmd stop
- service dcmd start