Retention policies and continuous queries

Telegraf
1

Hey everyone!

I’m new to InfluxDB and I need a little bit of help with continuous queries and retention policies…

What I’m trying to do is that I have one database in InfluxDB let’s call it “input” and another database which we can call “aggregated”. What I want is to aggregate and transport data from input database where they are ingesting every 0.5sec to aggregated database let’s say 1 datapoint for 30 sec., 1 datapoint every 20sec. and 1 datapoint every 10sec within a minute for example.

I have three CQs that look like:

cq_30sec CREATE CONTINUOUS QUERY cq_30sec ON input BEGIN SELECT mean(*) INTO aggregated.one_hour.cpu_load_short FROM input.autogen.cpu_load_short WHERE time < now() - 30s GROUP BY time(30s) END
cq_20sec CREATE CONTINUOUS QUERY cq_20sec ON input BEGIN SELECT mean(*) INTO aggregated.one_hour.cpu_load_short FROM input.autogen.cpu_load_short WHERE time < now() - 20s GROUP BY time(20s) END
cq_10sec CREATE CONTINUOUS QUERY cq_10sec ON input BEGIN SELECT mean(*) INTO aggregated.one_hour.cpu_load_short FROM input.autogen.cpu_load_short WHERE time < now() - 10s GROUP BY time(10s) END

I visualize it with Grafana where I have two graphs - first for input db where I have datapoints per 0.5sec and second one for aggregated where I have 1 datapoint per 10sec.

What I want to accomplish is to transport datapoints from iiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiii to --> i i i i i i i i i i i iiiiiii if we imagine that letters i are datapoints.

Let me know if is any further clarification needed.

Any help would be appreciated! Thank you for any reply!!

2

Yes, what you describe is very sound/valid, but which part isn’t working ?

Either the cq somehow isn’t running, is pointing from/to the wrong place, or new data (or maybe not enough new data) has been received yet. Unless I’ve misunderstood what the issue is?

3

AFAIK you don’t need to add the
WHERE time < now() - 30s
in continuous queries. That is added automatically by influx when the query is executed.

1 Like