Notification when value/row have been added

InfluxDB 2 Checks & Notifications
1

Dear all,

i am trying to achieve to get a notification if a new value (new row with a newer timestamp) have been added to a measurement. Is there any way to do this using a Task or a Check?

Thanks and BR,
Tobias

2

Could you provide a small example of the rows , and what you’re wanting to test/alert for?

3

@FixTestRepeat

Query:

from(bucket: "TestBucket")
  |> range(start: v.timeRangeStart, stop: v.timeRangeStop)
  |> filter(fn: (r) => r["_measurement"] == "movements")
  |> filter(fn: (r) => r["_field"] == "csv")

Here you see the result

Bildschirmfoto 2022-02-15 um 21.36.42
Bildschirmfoto 2022-02-15 um 21.36.422574Ă—544 107 KB

So, every time a new row (a new csv string) have been added I want to send a notification with some infos of that string. I could do this via a Telegraf Output plugin but am looking for a solution to directly use the InfluxDB own Notification/Alert system.

Thanks an BR,
Tobias

4

You honestly might be better off exploring those other alternatives. The two built in alert types are the threshold check and deadman check. If I understand your goal, you want to record all events and then sent a single summary alert of the latest “batch”.

If using influxdb for this, you would probably need to use a task that stores the last read/process event in its own table/measurement, and then collects all new events from this point in time onwards to now.

The built in alerts currently

Creating a set of tasks to achieve what you want:

5

So I finally found out how I can get this done.

option task = {name: "csv_detection", every: 10s}

lastDetected = from(bucket: "notifications")
    |> range(start: 0)
    |> filter(fn: (r) => r["_measurement"] == "movements" and r["sub_area"] == "T1")
    |> filter(fn: (r) => r["_field"] == "csv")
    |> last()
    |> findRecord(
        fn: (key) => true,
        idx: 0,
    )

from(bucket: "SourceBucket")
    |> range(start: if exists lastDetected._time then lastDetected._time else time(v: "1970-01-01T00:00:00Z"))
    |> filter(fn: (r) => r["_measurement"] == "movements" and r["sub_area"] == "T1")
    |> filter(fn: (r) => r["_field"] == "csv")
    |> to(bucket: "notifications")

This does the job absolutely perfect!

1 Like