I wrote a bash script for inserting the summary of the daily energy database "xxx_d" into the monthly database "xxx_m" - triggered on every 1st day of the month at 02 o'clock by a systemd-timer. That's a pretty nice working workaround for the missing "1M" on the continuous querys of InfluxQL.
Now I've the following issue (it's an example; every entry is timestamped too early): the entry in the monthly database "xxx_m" is written on 01.04.2023 02:20 with the wrong timestamp of the first data of the daily database "xxx_d", because this is the summary of all data from 01.03.2023 00:00 to 01.04.2023 00:00, but the timestamp is 01.03.2023 00:00 !
The problem: grafana shows the bar chart at 02-2023 - WHY? So all entries are one month too early. How can I solve this? Is the only working solution to make two separate influxQL-Commands from it? First a SELECT on the daily DB and then second an INSERT on the monthly db?
Until now it is a "One-liner" like where ${ANFANG} is "2023-03-01" and ${ENDE} is "2023-04-01".
SELECT sum(*) INTO xxx_m.autogen.:MEASUREMENT
FROM "shellies"
WHERE time >= '${ANFANG}' AND time < '${ENDE}'
Or is it a grafana issue and I need to solve it there? How to do that?
