Adding retention policy results in 400 error for feeds

I have Influxdb & Telegraf installed & running in a Oracle cloud instance.
The database was created automatically by Telegraf, and I can see feed data being successfully saved in Influx, everything looking good so far!

However, I want to introduce some retention policies, so have stopped Telegraf, and applied the following retention policy to Influx;

If I then check the policy, it all looks good;

name    duration shardGroupDuration replicaN default
----    -------- ------------------ -------- -------
autogen 0s       168h0m0s           1        false
one_day 24h0m0s  1h0m0s             1        true

But upon restarting Influx & Telegraf, I cannot retrieve any data from the database, and looking at the Influxdb log, I can see that the incoming data feeds are being rejected by Influx with a 400 code.

I’ve deleted the database & repeated the process with the same results.

Any advise please?

I’ve just checked Telegraf’s log, and I see;
2019-11-20T22:01:30Z W! [outputs.influxdb] when writing to []: received error partial write: points beyond retention policy dropped=2

Google’ing the error suggests;

“You’re trying to write a point to a retention policy that has a timestamp that’s so old, that the retention policy would have immediately scrubbed it from the database. For example, if your retention policy is 6 months (keep data for 6 months then delete it), and you try to add a point with a timestamp that’s 8 months in the past, then you would see this error.”

But in my case, the timestamp added to the feed value is a epoch (in milliseconds granularity) which is the same as 2019-11-20T22:01:30
precision = "ms"

Increasing the retention policy to 1 year, and the feed points are successfully saved in the database, so why is this working for a policy of 1 year and not for 1 day??