InfluxQL vs Flux performance on Influx 1.8

Fluxlang
, , , ,
1

I have been using both Flux and InfluxQL recently, comparing which is faster and how to query using both efficiently. I noticed a pretty significant difference in performance between Flux and InfluxQL and was wondering if there is an explanation for it.

This InfluxQL call on a 1.8 runs in less than a second:
SELECT last(“Value”) AS “mean_Value” FROM “MonitoringDB”.“autogen”.“Neurio.TotalConsuption.Power”

whereas this Flux call on a 1.8 server runs in ~10 seconds:
from(bucket: “MonitoringDB/autogen”)
|> range(start: 0)
|> filter(fn: (r) => r["_measurement"] == “Neurio.TotalConsuption.Power”)
|> filter(fn: (r) => r[“EngUnit”] == “Watts”)
|> filter(fn: (r) => r["_field"] == “Value”)
|> last()
|> yield(name: “last”)

They are essentially the same query. Is this something that has been fixed and is there an explanation for why this occurs?

2

Hello,

The performance of this kind of query has indeed been fixed in InfluxDB 2. In InfluxDB 2, the last function is covered by a pushdown, which makes the query run much faster when using flux and should be comparable to the InfluxQL query performance.

1 Like