You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Allow to set the precision of timestamp set by the input jti_openconfig_telemetry.
Currently timestamp precision is in ns. It could be interesting to have precision in second to avoid ms jitter between sensor paths.
Opening a feature request kicks off a discussion.
Proposal:
Allow to set :
precision: "s | ms | us | ns"
Current behavior:
Send ns timestamp
Desired behavior:
Still sending ns timestamp (epoch) but with less precision that a ns. Ex.
Before you have timestamp like that :
1557755590589715534
With precision="s" you should have something like :
1557755590000000000
Use case:
If I request to Junos from telegraf the sensor : /interfaces/interface
I could receive from the network equipment :
/interfaces/interface/state/counters/xxxx with timestamp 1557755590589715534
/interfaces/interface/state/cdescription with timestamp 1557755590685780881
To aggregate (merge) those 2 values in a third one on influxdb based on time it's more difficult.
It could help to have at the output of telegraf :
/interfaces/interface/state/counters/xxxx with timestamp 1557755590000000000
/interfaces/interface/state/cdescription with timestamp 1557755590000000000
The text was updated successfully, but these errors were encountered:
It may be worth adding a precision option for all input/aggregator plugins that would override the global agent precision. Unlike the agent precision it would apply to both "service inputs", like jti_openconfig_telemetry, and regular polling inputs.
Feature Request
Allow to set the precision of timestamp set by the input jti_openconfig_telemetry.
Currently timestamp precision is in ns. It could be interesting to have precision in second to avoid ms jitter between sensor paths.
Opening a feature request kicks off a discussion.
Proposal:
Allow to set :
precision: "s | ms | us | ns"
Current behavior:
Send ns timestamp
Desired behavior:
Still sending ns timestamp (epoch) but with less precision that a ns. Ex.
Before you have timestamp like that :
1557755590589715534
With precision="s" you should have something like :
1557755590000000000
Use case:
If I request to Junos from telegraf the sensor : /interfaces/interface
I could receive from the network equipment :
/interfaces/interface/state/counters/xxxx with timestamp 1557755590589715534
/interfaces/interface/state/cdescription with timestamp 1557755590685780881
To aggregate (merge) those 2 values in a third one on influxdb based on time it's more difficult.
It could help to have at the output of telegraf :
/interfaces/interface/state/counters/xxxx with timestamp 1557755590000000000
/interfaces/interface/state/cdescription with timestamp 1557755590000000000
The text was updated successfully, but these errors were encountered: