How to disable metrics averaging?

[makrushin-evgenii]

Hi. I run collector and monitor it with this Grafana dashboard. With 1 RPS and 10 RPS load, I got this metrics from query:

When the load instantly increased to 10, the graph went to this value for 5 minutes. When the load instantly disappeared, the graph also fell for 5 minutes

Looks like result of movingAverage(, 5min). How to get real load metrics?

# query from https://grafana.com/grafana/dashboards/15983-opentelemetry-collector/
sum(${metric:value}(otelcol_receiver_accepted_log_records${suffix}{receiver=~"$receiver",job="$job"}[$__rate_interval])) by (receiver $grouping)

What I tried?
Replacing all variables from query with constants (sum(rate(otelcol_receiver_accepted_log_records{receiver=~".*"}[10s])) by (receiver , )), i got this:

A way better, but not accurate. Graph show 10 records per 10 seconds, not 10 records per 1 second. Also, I dont understand why query results are different, if queries are equal

[makrushin-evgenii]

UPD $__rate_interval wasnt 10s. This explains the smoothing

sum(${metric:value}(otelcol_receiver_accepted_log_records${suffix}{receiver=~"$receiver",job="$job"}[10s])) by (receiver $grouping)

sum(${metric:value}(otelcol_receiver_accepted_log_records${suffix}{receiver=~"$receiver",job="$job"}[$__rate_interval])) by (receiver $grouping)

[makrushin-evgenii]

Noticed another oddity. At first I get a smooth metric, the holes appear a little later