Bob Steers

Note: I have little to no practical experience with PyPy. That said: This seems to be the primary influx + asyncio library. It may be desirable to focus on doing...

This also happens if the `host` property is set explicitly, and breaks with previous behaviour. The following snippet connects to 'remote-host' when using 4.1.0, and to 'localhost' when using 4.2.1...

Yes, I went through the suggestions listed in the troubleshooting page. To verify, I tried them again: - `query=avg_over_time({__name__="spark4/SystemInfo/uptime[second]"})` and `query={__name__="spark4/SystemInfo/uptime[second]"}` both reproduce the issue. - The time-dependent behavior make...

What additional info would be required? As mentioned before: I can provide the dataset I'm using to reproduce the issue.

The gzipped dataset was still 650MB, and too large to attach to this issue. I uploaded it to WeTransfer, and emailed the link along with the configuration/commands.

I'm aware of the gap. It's why I selected this particular field. The bug report is that if I query with a start date of 12:00 (before the gap), and...

It appears I made a typo in my original post: the query start date is `2022-10-03T15:00:00Z`, but should be `2022-10-03T12:00:00Z`. (15:00 -> 12:00) My apologies for the confusion. Otherwise, I...

Data ingestion does not involve backfilling. There's an MQTT listener that converts received messages to the Influx line protocol, and then immediately calls the `/write` endpoint. #3309 seems vaguely related,...

Update: I can confirm that the bug is still present in 1.83.1.

Here you go: https://we.tl/t-nSw3tUvCFp