Very slow v2.0.0-beta.16

InfluxDB 2
1

Hi,
is it recognized that v2 with flux it tremendously slower than v1 in fetching data?

There are two influxd servers running on same type of machine (AWS t2.xlarge) with same data pushed. Data is one series of 4Hz samples.

  1. v2.0.0-beta.16
from(bucket: "<bucket>")
  |> range(start: 2020-10-09T08:52:47Z, stop:2020-10-16T08:52:47Z)
  |> filter(fn: (r) =>
    r._measurement == "<measurement>" and
    r._field == "<field>"
  )    
  |> aggregateWindow(
    every: 5m0s,
    fn: mean
  )

takes 18seconds for fetch 2017 rows

  1. v1.8.3
SELECT mean(<field>) FROM <measurement> WHERE time >= now() - 7d GROUP BY time(5m) fill(null)

tasks 3.2seconds for fetching 2017 rows

Both process about 2M points. 3s vs 18s! How comes?

Peek 2020-10-16 11-07
Peek 2020-10-16 11-071847×649 144 KB

2

All of the beta releases have a note saying that they aren’t for performance testing just yet.

Also , the influx dB 2.0 beta releases have been replaced by rc0 , suggest upgrading to that. Note, that the database storage format is incompatible- but the time invested in upgrading now now should pay off in future.

1 Like