It doesn’t appear to survive a reboot (I’m having to re-enter the environment variables to theTelegraf environment and then run the start command.

I’ve tried setting up a influx-cli.service to run but it doesn’t …

Can anybody point me in the right direction to make this start on reboot?

Well, the header with the chosen bucket and linux_host disappears and all metrics show:

error @3:37-3:38: record is missing label linux_host
error @2:19-2:20: record is missing label timeRangeStart
error @2:43-2:44: record is missing label timeRangeStop
error @1:14-1:15: record is missing label bucket

If I do a ‘hard’ refresh through F5 everything gets back to normal.

Yes, you can. I have 2 pi’s running in 2 different internal networks. And influxdb is running on one of the 2. To be able to reach my influx-environment I’ve a subdomain influxdb.domein where it is pointed to my ‘given’ ip-address. Nginx then handles the connection to the influxdb-environment which is running on http://someip:8086.
To be able to get data from the second pi I use as INFLUX_HOST the influxdb.domein-value.

Thank you!

Already done.

Hi.
Thanks for sharing this.
Can you update the facebook group post with this solution?
Thanks.
Regards,
Sara

Solved it and posted as as response to a question on the RNT Facebook page).

Here’s what I wrote :

Out of the box, Telegraf starts automatically as a daemon with its default telegraf.service systemd script. It looks for a configuration file called telegraf.conf in /etc/telegraf/ to run. The default telegraf.conf file basically does nothing so just overwrite it.

Create a new telegraf.conf by saving your script from the UI :
Load Data -> TELEGRAF tab -> edit -> DOWNLOAD CONFIG

Note that the telegraf.conf file gets the values for $INFLUX_HOST, $INFLUX_TOKEN and $INFLUX_ORG from /etc/defaults/telegraf. So you need to have the correct values there. Fill it in something like this :

INFLUX_HOST=”http://localhost:8086″
INFLUX_TOKEN=” ”
INFLUX_ORG=” ”

or just edit them directly into your config file (Useful if you plan to run multiple instances of telegraf daemons).

HTH

While that works fine, it only sends data while that command is running. If you do a to get back to the command line that kills the second instance of telegraf and data about the pi stops getting sent to influxdb.

Do you have any hints about how to add raspberry pi monitoring to always run without needing a command line open? Possibly some changes to /etc/telegraf/telegraf.conf?

TIA

Hi.
Thanks for sharing your experience with us.
I’m not sure if that’s what caused the issue. I need to try with another token name and check what happens.
Regards,
Sara