Performance tuning to get DataFrame (335280, 21) in one second

Welcome & Getting Started
1

Dear ALL,

I’m new to InfluxDB. Right now, we have 10 years of data. We need to get one-day data (size: 200MB, 335280 rows, 21 columns) in less than one second. Right now, we are doing POC on the InfluxDB standalone. However, it can’t meet our targets yet. Would you please suggest the performance tuning advice/guides for us to speed up our query? We need nearly one minute to get one-day data at this moment.

Keep safe!

Best regards,
Yuyong

2

How about showing us the query you are using?

Then we might have a starting point to suggest what could be improved.

Also, which version of InfluxDB are you using, what operating system and
version is it installed on, and what sort of hardware is it on?

Antony.

3

Thanks, Antony!

Here is the python code I used.

client = DataFrameClient(host=bar_env.influxdbHost, port=bar_env.influxdbPort)
query = "select * from bardata.autogen.bardata where transactionDate = '2009-01-05';"
barDB = client.query(query)

InfluxDB version is 1.8.

The database is running on AWS Linux (4.14.177-139.254.amzn2.x86_64), 8CPU & 64G RAM, network bandwidth up to 25Gb