Hi,
I have a custom node exporter from which prometheus scrapes every 30
seconds.
Issue is that sometimes -
a) if the data volume is large ,node exporter takes longer , On top of that
,
b) i need to clear the metrics to remove the old stale values before i set
the metrics to the respective values again.
These scenario a) causes mixup of data from 2 iterations and in scenario
b) scrape cycle may get 0 values .
Expected behaviour -> Iteration 1 starts -> Node exporter collects raw
data -> clear metrics -> process data and set metrics -->Iteration 1
complete -> Prometheus Scrapes Data.
Problem is : a) Iteration 1 starts -> Node exporter collects raw data ->
clear metrics -->* ( Prometheus Scrapes Data ) *-> process data and set
metric #1 --> . process and set metric #N -->Iteration 1 complete
b) Iteration1 starts -> Node exporter collects raw data -> clear metrics
--> process data and set metric#1 --> *( Prometheus Scrapes Data ) * -->..
process and set metric#n--> Iteration 1 complete
So, Is there a way to enforce some syncronization between node exporter and
prometheus scrape cycle ? example-
Scenario 1) The node exporter prepared data in 20 seconds and Prometheus
should start scraping from ( say) 30th second of every minute i.e.
02:34:30
Scenario 2) To expose all the metrics in one shot , example -
Iteration2 starts -> Node exporter collects raw data -> clear metrics -->
process
data and set metric#1 --> -->.. process and set metric#n--> Iteration 2
complete -> expose metric.
something like - metric1.labels(STATE=x).set (y)
metric1.expose()
I am okay to scrape old data , but dont want to -
mix Iteration 1 and Iteration 2's data or ,
scrape the data when node exporter has cleared metric.
Scenario 3: make scrape wait till data collection iteration is complete.
Please do let me know if more information is required from my end here.
Regards
Puneet
--
You received this message because you are subscribed to the Google Groups
"Prometheus Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/prometheus-users/bc673f27-dcf3-4cf0-a190-2c2cb537b259n%40googlegroups.com.
