time client_ip client_port host path pid program syslog_server syslog_timestamp
1524129916761728248 10.1.36.131 53842 thw-dv-lamp-01 /data/mnt2/haproxy-2018-04-17_12_13.log 11735 haproxy 10.1.36.10 Apr 15 22:59:59
Is time Micronsecond?
time client_ip client_port host path pid program syslog_server syslog_timestamp
1524129916761728248 10.1.36.131 53842 thw-dv-lamp-01 /data/mnt2/haproxy-2018-04-17_12_13.log 11735 haproxy 10.1.36.10 Apr 15 22:59:59
Is time Micronsecond?
Looks like nanoseconds, you can change the display precision in the influx shell using the precision
command:
precision <format> specifies the format of the timestamp: rfc3339, h, m, s, ms, u or ns
Thanks a lot. I confirm that the default is indeed nano-seconds
Hi Daniel,
I had added precision = “s” to my simple telegraf config:
[[outputs.influxdb]]
precision = “s”
However, it seamed not make any difference - i still have ns in my influxdb
Please shed some light.
Thanks a lot in advance
InfluxDB always stores in nanosecond precision internally, so sending this option only rounds to the nearest second. However, this option hasn’t done anything for quite a few versions of Telegraf, instead you can influence the rounding of most inputs by setting the precision option in the agent:
[agent]
precision = "" # default is 1s
This doesn’t apply to “Service Inputs” which are the inputs that are event driven instead of interval driven, examples are socket_listener or statsd.
Thanks a lot indeed. Shall try out soon