Hi all
I’m extracting resources from 4 different servers using telegraf agents and then, send the outputs to host server’s influxdb. (interval = 5s)
The thing is… the time!!
The time of 3 servers is exactly same each other, but the last one’s time is slightly different(only 1 sec) like below
the servers’ date seems to be same, so I don’t know why they’re different.
Can I solve this problem via config or sth?
I appreciate your response in advance.
This is known to be possible, double check that you have round_interval enabled in your agent configuration, though I bet it is is already set.
By default Telegraf rounds to the nearest second from the actual time of collection, and it must be collecting a little early on one system. For most use cases this isn’t much of a problem, and you should just ignore it, but if its very important to have the times line up you could set precision = "5s" in the agent configuration.
We have some upcoming changes to the scheduler which will make this occur less often, but it will still be possible since the reported time is based on actual time, not scheduled time.
A few ideas…
Are all the systems using the same Network Time Protocol servers?, if not there may be discrepancies.
How accurate do you need the time to be?
Are the servers located in the same building?
Have you considered a local hardware Network Time Protocol server?
I’m sorry, I’m late.
It works very well!
Thank you. You saved me.
