The chain from Telegraf (influxdb_v2) to InfluxDB (cloud) seems to silently drop measurements that have non-ascii in name or tags

I have been evaluating Telegraf and cloud InfluxDB to see if they would fit my needs. For this I have set up a simulation with a python script that feeds simple line protocol entries via http to Telegraf influxdb_v2_listener. Telegraf uses influxdb_v2 plugin to send those to InfluxDB cloud account.

This works fine as long as I keep the measurement name, tag names and values as ascii. The measurements get to InfluxDB. But when I include any of the scandinavian umlaut characters like åäö in the name or tags it stops working. The Telegraf debug log shows that is has sent those successfully to InfluxDB but they do not show up in InfluxDB.

Is this a known issue?

Telegraf is running locally in docker on my Windows 10 PC in Ubuntu 18.04 in WSL. Telegraf is the latest docker image as of today. My python script runs locally in the same Ubuntu setup.
The beginning of the Telegraf log is below. It is showing that all is loaded just fine

2020-11-22T17:10:47Z I! Starting Telegraf 1.16.2
2020-11-22T17:10:47Z I! Using config file: /etc/telegraf/telegraf.conf
2020-11-22T17:10:47Z I! Loaded inputs: influxdb_v2_listener
2020-11-22T17:10:47Z I! Loaded aggregators:
2020-11-22T17:10:47Z I! Loaded processors:
2020-11-22T17:10:47Z I! Loaded outputs: influxdb_v2
2020-11-22T17:10:47Z I! Tags enabled: host=testpc
2020-11-22T17:10:47Z I! [agent] Config: Interval:10s, Quiet:false, Hostname:"testpc", Flush Interval:10s
2020-11-22T17:10:47Z D! [agent] Initializing plugins
2020-11-22T17:10:47Z D! [agent] Connecting outputs
2020-11-22T17:10:47Z D! [agent] Attempting connection to [outputs.influxdb_v2]
2020-11-22T17:10:47Z D! [agent] Successfully connected to outputs.influxdb_v2
2020-11-22T17:10:47Z D! [agent] Starting service inputs
2020-11-22T17:10:47Z I! [inputs.influxdb_v2_listener] Started HTTP listener service on :8086
2020-11-22T17:10:57Z D! [outputs.influxdb_v2] Buffer fullness: 0 / 10000 metrics`

When the python script sends a line protocol entry to Telegraf like
cabin,from=tester,bucket=testbucket,house=this pulse=1 1606070051092467500
the measurement gets through and is available in InfluxDB. Telegraf shows in the log the following.
2020-11-22T18:34:18Z D! [outputs.influxdb_v2] Wrote batch of 1 metrics in 418.3693ms

If i change the measurement name ot the tags to have non-ascii characters everything looks the same except that the measurement is not showing in InfluxDB.
mökki,from=tester,bucket=testbucket,house=this pulse=1 1606070235475961800
2020-11-22T18:37:18Z D! [outputs.influxdb_v2] Wrote batch of 1 metrics in 111.9464ms

I can see that InfluxDb can handle the non-ascii entry since I can enter it manually via the Load Data->Buckets->Add Data->Line Protocol->Enter Manually dialog. It will happily accept the entry
mökki,from=tester,bucket=testbucket,house=this pulse=1 1606070235475961800
and the entry will show up in Explore.

My Telegraf configuration has nothing funny in it, just the generated one for these two plugins with my credentials added.

Would this perhaps be a bug in influxdb_v2 plugin or InfluxDB itself?