I'm creating alerts in Prometheus and migrating from Datadog.
I have two metrics queries that I'm not able to understand yet.
avg(last_1d):anomalies(avg:default.burrow_kafka_consumer_lag_total{*} by {consumer_group,env}, 'robust', 3, direction='above', alert_window='last_30m', interval=300, count_default_zero='true', seasonality='hourly') >= 1
In this query I understand the avg:default.burrow_kafka_consumer_lag_total{*} by {consumer_group,env} part but not the rest and how to translate it to PromQL.
Second
min(last_2h):derivative(avg:default.burrow_kafka_consumer_lag_total{!consumer_group:connect-analytics-mobile-json} by {env,consumer_group}.rollup(avg, 300)) > 0
I don't understand the rollup part, how to translate it into PromQL?
pct_change(avg(last_1h),last_1h):avg:default.burrow_kafka_consumer_lag_total{!consumer_group:vdv_trip_aggregator_app,!consumer_group:vdv_trip_normalizer_app-fast,!consumer_group:vdv_trip_normalizer_app-slow,!consumer_group:disruption-consumer-app} by {consumer_group,env} > 300
how is pct_change(avg(last_1h),last_1h) part of the query?
I'm new to this. I have translated other queries but these I don't understand.
