scrape_interval is not working properly.

[Byungkwon Choi]

Hello,

I want to collect the HTTP requests per second every second.

To do so, I'm using Prometheus and Prometheus Adapter.

I set the value of 'scrape_interval', 'scrape_timeout', and 'evaluation_interval' to 1s in the Prometheus ConfigMaps resource of K8S as follows:

I use the following query to collect HTTP requests per second:

sum(irate(request_total{namespace!="", direction="inbound"}[30s])) by (deployment)

However, the value of the query is updated almost every 20 seconds.

What I expect to this value is to be updated every second because I use 'irate' here and set 'scrape_interval' to 1 second.

My current environment is as follows:

• Service mesh: Linkerd v2.8.1
• K8S version: v1.17.9
• CNI: Calico
• Other components deployed together: Jaeger, Prometheus-adapter

Did I miss something?

Thank you in advance!

[Ben Kochie]

You didn't provide any evidence of why you think it's not working. You need to include more information.

--
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 prometheus-use...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/prometheus-users/aa2fcdf9-dddc-4c36-b1be-82cd9bae8542o%40googlegroups.com.