Hi,
I’ve setup influxDB for my time series data to log total consumption counters. This was a succes by far. Now that I have the data I want to visualize it as follow:
RAW DATA as follow:
select value from consumption
name: consumption
time value
2018-05-01T09:37:56Z 66
2018-05-02T10:18:44Z 112
2018-05-03T08:07:13Z 142
2018-05-05T09:17:25Z 258
2018-05-06T08:03:01Z 336
2018-05-07T08:01:13Z 378
2018-05-08T08:34:12Z 442
2018-05-09T08:44:08Z 498
2018-05-10T07:25:08Z 541
2018-05-10T07:35:28Z 542
2018-05-10T07:55:28Z 542
2018-05-10T07:59:55Z 543
2018-05-10T08:01:55Z 544
2018-05-11T08:23:21Z 593
2018-05-12T09:01:21Z 659
As you can see, the data looks like daily received, but definitely not perfect on the same time, making the difference() function nog suitable for this.
Also, notice the data loss on 2018-05-04
Also, notice the excess of data on 2018-05-10
These 3 facts that can and will happen with the data make the functions difference() and integral() etc… not suitable for error prove visualization.
also, it is possible that meter1 makes his data available every day, meter2 every hour, meter3 every 15 minutes, and meter4 not time based but event based of threshold based.
I want to visualize this data as consumption within variable time blocks, let say, all meter consumption on 15 minute base.
Does anyone knows the solution or a hint for that?
I allready searched google, grafana and influx forums. Wat terms I allready found are pre-bucketed data, and that the histogram() function for influxdb is not available yet.
I reconstructed the problem on my homeserver at:
http://81.83.6.53:3000
I created demo data for meter1 and meter2. This is the printscreen from meter2, with 2 times missing data and once excess data. With the difference() function I get 2 spikes and one fall.
