Hi there guys,
So I have been trying to get streaming telemetry going but I have encountered numerous issues.
The biggest problem is that the timestamp is defined in nanoseconds.
So this also means that I cannot match together meassurements and for example look at descriptions on interfaces and then group on that in influxdb.
I am currently running Telegraf 1.17.2
My current sensor configuration looks like this:
sensors = [
"2000ms interfaces /interfaces/interface/",
"2000ms system /system/"
]
Then i've got some processor renames which makes the fields abit easier to select on.
From reading from the juniper documentation; i have seen that they recommend that you aggregate data points becuase of the fact that the data can sometimes arrive at uneven intervals which then causes my group_by timeinterval in grafana to display very unrealistic values.
Flattening would be nice if it were possible to achieve wihtout having to write my own tools for that, has anyone been able to solve this?
This is an example on this:
Most of my hardware is QFX5k devices.
name: interfaces
time interface_name description in-octets out-octets
---- -------------- ----------- --------- ----------
1615479803310511836 et-0/0/0 MY_SUPER_DESCRIPTION_OF_DOOM
1615479804051760418 et-0/0/0 11011600237580 10968460462666
I cannot retrieve in and out-octets based on description.
It would be very interesting to see how/if anyone else has had the same issue
This is a desired result
name: interfaces
time interface_name description in-octets out-octets
---- -------------- ----------- --------- ----------
1615479803310511836 et-0/0/0 MY_SUPER_DESCRIPTION_OF_DOOM 11011600237580 10968460462666
------------------------------
Oskar Hammarstrom
------------------------------