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
{{ message }}
This repository was archived by the owner on Aug 4, 2021. It is now read-only.
I am not 100% sure that what I will write is completely true, but we are using the pg_prometheus with normalized storage for some time now. Today we wanted to create some basic continuous aggregates view so that our dashboards are more responsive when looking to the long-term data.
Unfortunately we constantly fail to do so. Even the most simply aggregate like:
prometheus=# CREATE VIEW basic
prometheus-# WITH (timescaledb.continuous)
prometheus-# AS
prometheus-# SELECT
prometheus-# time_bucket('1 day', time),
prometheus-# sum(value) AS "total"
prometheus-# FROM metrics
prometheus-# GROUP BY time_bucket('1 day', time);
ERROR: invalid SELECT query for continuous aggregate
fails with the invalid select query error message. The select query itself works without any problem of course.
As I understand this, the creation most likely fails as the metric is not a table, but a view. If this is the case, how can we perform continuous aggregates with the normalized storage? Would I need to go directly to the metrics_labels and metrics_values tables?