Line Protocol Timestamp seems off 1 hour?

InfluxDB 2
1

Hi,

First time on the community with a question. During experimenting, where everything worked smooth and fine, I just recently noticed something.

I’m having an issue with the use of the Line Protocol that I am trying to solve.

Running v2.7.6 on a local linux machine.
Tried the REST WS and the Line Protocol “Enter Manually” option.

profiles,name=GGV,commodity=GAS azi=0.00016932528 1577836800
profiles,name=GGV,commodity=GAS azi=0.00020000000 1577840400

Deliberately using the timestamp from the influxDB docs site.
https://docs.influxdata.com/influxdb/v2/reference/glossary/#unix-timestamp

It should correspond with 2020-01-01T00:00:00Z UTC datetime.

But when I’m querying it seems to have stored it under
_time timestamp 2020-01-01T01:00:00:00Z.

influx2
influx21586×185 24.3 KB

Might there be something I’ve missed or doing wrong here?
Maybe a server setting of any kind?

Thanks in advance,

Roeland

2

Hi,

I believe I’ve got the answer to this. I think I’ve seen this issue explained before a while back on this forum.

While writing my data with a regular interval every hour it depends on how you query this interval.

|> aggregateWindow(every: 1h, fn: mean, createEmpty: false)

Maybe someone can confirm but I can understand that an aggregate of 2 timestamps will write the result into the second timestamp. I just did not expect it would also happen when my interval is already 1 hour.

Does anyone know what would be the correct way to resolve this?
Either I remove the aggregateWindow or I expand my aggregateWindow with a additional paramter.

timeSrc: "_start"

This addition seems to write the value at the first timestamp though I am not sure what this option does exactly.

Does anyone have a suggestion?

Regards,

Roeland

3

By default, aggregateWindow() uses the _stop column (the stop time of each window) to get the new timestamp for the aggregated point.

I.e. mean value from 10:00 to 11:00 would have 11:00 timestamp.

Setting timeSrc to _start you change this behaviour and get 10:00 for an example above.

Without using aggregation at all you are viewing raw data that should be the same as in your line protocol.

4

Hi ebabeshko,

Thanks for the response.

And yes, this definitely confirms my findings. Now I can continu from here.

Cheers,

Roeland