Skip to main content
Question

Grafana dashboards -- custome query -- aggregate creates not valid datapoints

  • October 6, 2025
  • 2 replies
  • 41 views

 

Below I have attached three screeshots where I query same data in same time range. I have used Queries in first screenshot, and it corresponds to original data inspected in Charts, i.e. there is no data before 4:01. This is very basic query using external id. However in many cases we need custome query where we use parametrised syntax. I have used a very basic syntax in the second screenshot that only use external id to query same timeseries as first screenshot. But the result is clearly different, there are datapoints created where there should be no data. Inspecting the datapoints in query, I see real data points with some values are created in no data area! Disabling aggregate resolve problem of creating non exiting data in third screenshot but then it query all datapoints and saturates in 2 seconds (reach max datapoints that grafana can show). So I need aggregated data and I have tried aggregate inside the syntax and same problem in scheenshot2 persists in this case.

 

Any idea how to overcome this issue?

 

 

 

 

 

2 replies

ipolomanyi
Practitioner
  • Practitioner
  • October 7, 2025

Custom query uses synthetic timeseries query for data fetching. It interpolates data points for the sparse data: https://developer.cognite.com/dev/concepts/resource_types/synthetic_timeseries.html#interpolation

I am not aware of any simple workaround for this, except:
1. referencing a single time series directly (what you’ve done)
2. leveraging Data Modelling tab to filter the relevant time series
 


  • Author
  • Committed
  • October 7, 2025

Custom query is crucial for us to make dynamic and efficient dashboards in Grafana. And it is a major quality issue if our clients finds datapoint where it supposed to be no data. I hope Cognite Team can find a solution for this.

Is there any reason why interpolation cant’t be avoided in synthetic timeseries?