Acquire heterogeneous data with telegraf via TCP

Telegraf
1

Hi all,

I have a device that sends a heterogeneous byte stream over TCP.
The goal is to write these data in the right format into influxDB.

The data I receive is in the following format:

Byte length Type
0 1 UINT8
1 1 char
2 1 char
3 1 char
4 1 char
5 1 char
6 1 char
7 1 char
8 1 char
9 1 char
10 1 char
11 1 char
12 1 char
13 1 char
14 1 char
15 1 char
16 1 UINT8
17 2 INT16
19 2 INT16
21 2 UINT16(16 boolean values)
23 1 char
24 1 char
25 1 char
26 1 UINT8(8 boolean values)
27 1 UINT8(8 boolean values)
28 4 UINT32
32 4 UINT32
36 4 UINT32

I’m trying to use the plugin socket_listener but I can’t figure out which format is suitable for my case

Is it possible to read these different data directly with telegraf? or is it better to create an interface software?
I ask so as to be directed in the right way.

2

If you send the data via UDP, you might directly use the binary parser to dissect the data. The `binary
parser is only available in master and nightly builds but is not yet released IIRC.
If you are not using UDP and your message isn’t terminated by a newline, you additionally need to use a binary found in PR #12187, Let me know if you have any issues!

3

Hi Srebhan,
Please excuse the delay, I noticed that binary parser is released now.

What Filename I should use if the data come to me via TCP?
image

4

@ATeam if data is sent via TCP you need to use the socket_listener input plugin to create a TCP server on Telegraf side.

5

It’s lack of good example and document, it’s go this way. Any complete example appeared in the document would be great help for who is new to influxdb.

[[inputs.socket_listener]]
service_address = “tcp://:2020”
splitting_strategy = “newline”
data_format = “binary”
endianess = “be”

[[inputs.socket_listener.binary]]
metric_name = “some name as measurement”

entries = [
  { bits = 32, omit = true },
  { name = "VoltageA", type = "uint16" },
    ]
[inputs.socket_listener.binary.filter]
   selection = [{ offset = 0, bits = 8, match = "0xA2" }]
6

I believe @srebhan was involved heavily on the binary parser. and kudos to him.

but I do agree that the examples and documentation were hard to digest regardless of that it’s a very nice feature.

7

@DanielX and @fercasjr it’s usually hard to write good documentation if you wrote the code. Please open a PR and extend the documentation. Also feel free to add new examples, best would be as test-cases so we also increase code coverage…