I currently have an InfluxDB Server receiving data from a number of sources on a daily basis (1 point per day per field). I setup a Check to indicate when some of these values go beyond certain key threshold values, but the checks don’t seem to be firing correctly despite the “Task ran successfully” status on the check itself (e.g. the field is above the threshold to set its status to Warning, yet it has not changed).
I have previously configured a check when receiving data at a much higher frequency, e.g. 10Hz, with a 10s check. Is it possible that a 1-day sampling rate is causing issues for the check (i.e. not seeing the data)?
That’s a very large offset. Perhaps the task actually hasn’t run yet? Because there’s an 8h delay in execution. I’d try a much smaller offset like 5 min.
If I run a check with 24h intervals, when will it execute? At 0h00 utc? The reason I included the offset is that I have only a single point in the 24h period, and I want to make sure that point has been populated before the check is run.
Perhaps I misunderstand how the offset works…
Offset doesn’t won’t change the range for which the data is being queried. It will only delay the execution of the task. If you don’t include an offset it will run at 0h00 but you should include an offset to avoid read and write conflicts. Unless you’re writing a point at 36:00 with a timestamp at 12:00 (I can’t think of a usecase for that so if you are i’m curious to learn about it!) you should be good to go.
Funny you should mention, that is pretty much what we are doing.
We are receiving data from a client site in batch (24h of data from 0h00-23h59 dumped on our servers via SFTP at 2h00). We then crunch some KPIs from this data and write a single KPI value on our influx server at 6h00 (for a number of reasons, we delay this a few hours from when we receive the data).
Effectively, we have a single value pushed to the InfluxDB Server at around 6h00 with a timestamp of 00h30