Whats the best way to add a new continuous query to a production db influxdb? Clone the production influxdb ? I was hoping to avoid that. Is there a good way to test and try them it just through the web admin interface?
2 Answers
I guess you could:
create a temporary measurement .. setup the CQ .. insert some sample data and check the result once the CQ executes
[better approach] Send your timeseries data to 2 different measurement tables in parallel. Run your CQ experiments on one of them and discard it once you have the query you like.
The web interface might not help you much because CQ don't return results when you execute. They just pipe the output to a new measurement table you define. You can then run queries on the down sampled measurement table once at least 1 iteration of CQ finishes successfully.
Would love to hear other thoughts though.

- 1,014
- 11
- 28
-
1In the end. I went actually with option two. I wasnt sure about the performamce implications. Since it was prod, i wanted to have a better idea, but i couldnt see any other way. – JE42 Nov 06 '15 at 07:53
New version of InfluxDB supports INTO
clause:
https://docs.influxdata.com/influxdb/v1.4/query_language/data_exploration/#the-into-clause
So you can run your own CQ SELECT
statement directly (don't forget add WHERE time
clause for GROUP BY time
).

- 413
- 3
- 10