The Server Density agent makes use of your server time to determine the timestamp for your metrics. When metrics are collected the current system time is included within the metric payload and Server Density uses this data to determine the timestamp of the metric when processing.
As Server Density uses the time from your server it is important to ensure that it is correct and in sync. We recommend installing and configuring an NTP daemon on your server to ensure that your clock is always in sync and to prevent any of the following possible issues.
Potential Issues Caused by Out Sync Clocks
If your server clock is out of sync you may see the following issues:
- Incorrect times on your metric graphs.
- Incorrect times in your metrics API responses
- Alerts not triggering, if the metric timestamp is more than 5mins in the past at the time Server Density receives the metric for processing.
Monitoring NTP
You can use Server Density to monitor your server time compared to an NTP server and alert on and graph clock drift. To get setup simply follow the Monitoring NTP instructions. Once the plugin is installed you can then alert on and graph the ntp.offset metric. We recommend alerting if the offset is more than 60 - If you have installed and configured an NTP daemon this shouldn't occur, however, creating this alert ensures that Server Density will notify you if there are any major ongoing issues with time synchronisation on your server.
Comments