Telegraf seems to be ignoring CSV headers

Telegraf
1

I’m trying to write to Influx using Telegraf by writing lines of CSV which are picked up by using Python. I’m having trouble because the headers that I’m specifying don’t seem to be be being used when writing to the database.

Probably easiest if I explain with an example:

meter = "something"
col_1 = 1
col_2 = 2
col_3 = 3
col_4 = 4

message_1 = ("meter,col_1,col_2\r\n%s,%0i,%0i" % (meter,col_1,col_2))
reporter.send_data(message_1)

message_2 = ("meter,col_3,col_4\r\n%s,%0i,%0i" % (meter,col_3,col_4))
reporter.send_data(message_2)

The ‘reporter’ function is the bit that sends via a Unix socket.

This is what the database looks like afterwards…

time                col_1 col_2 host meter
----                ----- ----- ---- -----
1651177687434054435 1     2     pi1  something
1651177687434422446 3     4     pi1  something

However I was expecting it to look more like this, with the second message populating col_3 and col_4 based on the names I used…

time                col_1 col_2 col_3 col_4 host meter
----                ----- ----- ----- ----- ---- -----
1651177687434054435 1     2                 pi1  something
1651177687434422446             3     4     pi1  something

Here’s the relevant bit of telegraf.conf…

[[inputs.socket_listener]]
    service_address = "unix:///tmp/telegraf.sock"
    data_format = "csv"
    csv_header_row_count = 1
    name_override = "battpack"
    data_type = "float"
    csv_tag_columns = ["meter"]

Just starting out with this so hope it’s not too dumb a question, but it’s stumped me all afternoon. Any ideas what I’m doing wrong?

Dave

2

Hi,

I would need to confirm this, but I don’t believe the headers are going to get processed for each message. Your first message had col_1 and col_2 defined, and so those will be used going forward. That tends to line up with your result.

You would want to have all your headers available from the first message.

2 Likes