Stop to load csv data during specific time

Telegraf
1

Hello Experts.

Please give me your helps.

On one measurement, I set up to load csv files. At first, the files were loading and saving in influxDB. But suddenly, the loading was stopping to load during specific time.

The loading is working during 0:00 - 8:30 (JST). On other time (8:31 - 23:59 (JST)), the loading is stopping. There is no error message on /var/log/messages and other measurement is working fine. It seems influxDB process is working fine. But influxDB is not loading these csv files.

And this issue was started from July 10. There were no issues between July 6 - 9. And now, it’s keeping not to load a data expecting 8:31 - 23:59 (JST), every day.

I might think it seems there is any scheduler for loading. But of cause, I don’t find a scheduler looks like to pause loading a data.

The loading is configured on “telegraf” configuration.

best regards

Yuki

2

you mention InfluxDB process is working fine, what about telegraf?

Telegraf service collects the data as you specified in the configuration file.
InfluxDB is the database.

so, if the collection is the problem, sounds like Telegraf issue, however I would like to ask you first.

If everything is running, have you validated that the missing csv data has the correct format?

maybe the source file is missing something?.

3

To add to @fercasjr great questions, if you do think this is an issue with telegraf, please provide the telegraf logs and your config. If you see any errors about parsing going bad an example line or data would be helpful as well.

What I would do first, is go look at the influxdb logs around the time of those failures as well and see what those logs show.

4

bro how did you send the csv data to influxdb

5

Hello Experts

Thank you for your all comments.

I understood your comments. But unfortunately, on same telegraf configuration, it’s loading snmp polling data. It’s working fine for snmp. And there is no specific / special error both influxdb and telegraf on /var/log/messages around starting time (0:00) and stopping time (8:30).

Is it need to collect telegraf debug logs? This is customer’s system. It’s difficult to collect any logs without customer’s agreements. If it’s easy way to collect telegraf log (for ex, cli command), please let me know.

This is part of telegraf configuration.

One csv file is transferred from other server each 5 min. And the file was saved on one directory. All of csv files are same file format and saved on one directory. And it’s roughly same file size between all files.

[agent]
interval = “300s”

[[outputs.influxdb]]
urls = [“http://127.0.0.1:8086”]
database = “telegraf”
timeout = “5s”
user_agent = “telegraf”

[[inputs.snmp]]
agents = [ “IP, IP”]
version = 2
community = “XX”
name = “snmp”

[[inputs.file]]
files = [“/var/data/*.csv”]

name_override = “csv”

data_format = “csv”
csv_header_row_count = 1
csv_delimiter = “,”
csv_skip_rows = 0
csv_skip_columns = 0
csv_comment = “”
csv_trim_space = false

csv_tag_columns = [“HOSTNAME”]
csv_measurement_column = “”

csv_timestamp_column = “DATE_TIME_ID”
csv_timestamp_format = “2006-01-02 15:04:05”

csv_timezone = “Asia/Tokyo”
csv_skip_values =

I might think the file doesn’t have any issues. Because if the file format has any issue, I don’t sure why loading was fine during partial time (0:00 - 8:30).

On 7/20, the loading was working during 23:55 (7/19) to 8:50 (7/20), JST…

best regards

Yuki