Unable to view any metrics for certain VMs
Unable to view certain metrics for certain VMs affecting observability and tracking of VMs in vROPS
In the ESXi bundle in /var/run/log/vobd.log we may see the following:
vobd.log
1149:2024-06-11T05:01:11.250Z: [ClockCorrelator] 7826272852729us: [vob.clock.correction.adjtime.sync] system clock synchronized to upstream time servers
vCenter 7.x
ESXi 7.x
Stats collection depends on the machine clock. The clock has a very important role as all aggregations, summations, or any other type of processing of a value depends on it. By design, until ESXi 8.x, when the clock moves, the samples are treated as not valid and the file which holds them is truncated.
If the clock moves with more than 1 collection time interval (20 seconds), the data collection re-starts successfully but the dynamic counters (all counters coming from StatsRegistry (guest counters are apart of it) are not collected.
Run this command on the ESXi host to see if clock movement is causing the issue:
grep -i 'system clock synchronized to upstream' /var/run/log/vobd.log | wc -l
Restart hostd on the affected hosts to remediate the issue:
/etc/init.d/hostd restart
Issue will be fixed in future releases.