InfluxDB source: Data is missing a time field

Grafana InfluxDB
1
  • Grafana 12.0.2, installed as container from image grafana/grafana-oss:latest on macOS
  • InfluxDB 2.7, installed as container from image influxdb:2.7.

Hi,

Being a total n00b with Grafana, please be forgiven about the level of this question. I try to integrate Grafana visualization with the open source project TSDuck and try to educate myself about Grafana.

I start with simple things, before getting more ambitious. I did a Grafana + InfluxDB setup using a dual-container and docker-compose, as documented online (exact procedure here if needed).

I get error “Data is missing a time field” trying to create a Flux query. The data do have a timestamp field. The problem is probably obvious but I couldn’t find an explanation.

Monitoring data are sent by the application to InfluxDB in this format:

bitrate,scope=ts,tsid=4 value=23722592 1750858324051
bitrate,scope=pid,pid=8191 value=0 1750858324051

I select “Create dashboard”, “Add visualization”, select “influxdb (default)” as data source, enter a Flux query such as:

from(bucket: "demo-bucket")
  |> range(start: -5m)
  |> filter(fn: (r) => r._measurement == "bitrate")
  |> filter(fn: (r) => r.scope == "ts")
  |> filter(fn: (r) => r.tsid == "4")
  |> mean()

After hitting “Refresh”, I get “Data is missing a time field”.

Here is the content of the InfluxDB database (shrinked view with _start and _stop time removed);

$ influx query 'from(bucket:"demo-bucket") |> range(start:-20s)'
Result: _result
Table: keys: [_start, _stop, _field, _measurement, pid, scope]
_field:string  _measurement:string    pid:string  scope:string  _time:time  _value:float
-------------  -------------------  ------------  ------------  ----------  ------------
        value              bitrate          8191           pid  ...000000Z             0
        value              bitrate          8191           pid  ...000000Z             0
        value              bitrate          8191           pid  ...000000Z             0
        value              bitrate          8191           pid  ...000000Z             0
        value              bitrate          8191           pid  ...000000Z             0
        value              bitrate          8191           pid  ...000000Z             0
        value              bitrate          8191           pid  ...000000Z             0
        value              bitrate          8191           pid  ...000000Z             0
Table: keys: [_start, _stop, _field, _measurement, scope, tsid]
_field:string  _measurement:string  scope:string   tsid:string  _time:time  _value:float
-------------  -------------------  ------------  ------------  ----------  ------------
        value              bitrate            ts             4  ...000000Z      23549632
        value              bitrate            ts             4  ...000000Z      23758430
        value              bitrate            ts             4  ...000000Z      23736128
        value              bitrate            ts             4  ...000000Z      23735855
        value              bitrate            ts             4  ...000000Z      23759888
        value              bitrate            ts             4  ...000000Z      23720271
        value              bitrate            ts             4  ...000000Z      23752360
        value              bitrate            ts             4  ...000000Z      23725600

Thanks for your help.

2

Welcome to forum @lelegard

Excellent way of providing a solid question! So your mean() function is returning just one value and if you are using time series visualization, it will complain about missing time field

image
image1504Ă—654 50.2 KB

try this instead

from(bucket: "tsduck")
  |> range(start: v.timeRangeStart, stop: v.timeRangeStop)
  |> filter(fn: (r) => r["_measurement"] == "bitrate")
  |> filter(fn: (r) => r.scope == "ts")
  |> filter(fn: (r) => r.tsid == "4")
  |> aggregateWindow(every: v.windowPeriod, fn: mean, createEmpty: false)
  |> yield(name: "mean")
2 Likes
3

Thanks @yosiasz, that’s much better !

I will read the docs to understand each part of your request and try to build more ambitious reports.

One final question. What is the difference between r._measurement (my original request) and r["_measurement"] (your modified request)? Especially because you kept the dot notation in other fields such as r.scope.

4

The difference lies in the syntax for accessing record properties:

  • r._measurement (Dot Notation): This is the more common and generally preferred way to access properties of a record in Flux. It’s concise and easy to read, similar to accessing properties in many other programming languages. You use dot notation when the property name is a valid identifier (i.e., it doesn’t contain spaces, special characters, or start with a number).
  • r["_measurement"] (Bracket Notation): This notation is used when the property name contains special characters (like _ in _measurement, _field, _value, etc., although dot notation does work for these standard InfluxDB columns).
    • Use bracket notation to access keys with special or whitespace characters.

In the specific case of _measurement (and _field, _value, _time, _start, _stop), both dot notation and bracket notation are valid and produce the same result. InfluxDB’s built-in columns are designed to work smoothly with dot notation despite their leading underscores.

2 Likes
5

Thank you @grant2, this is clear.