Search code examples
influxdbgrafana

Why has Grafana stopped to plot the newly written values from my influxdb database?

I have an influxDB database called "test0" of several variables stored every 30s on a local server on a raspberry pi with raspbian. I use this system to log data since several months without any problem. But since some days, Grafana do not plot the data newly written in the database.

Here is a zoom on the problem : in the CLI of influxDB, when I type the following query to get the temperature in a time range:

> SELECT Temp_cuve FROM test0 WHERE time > now() - 19h53m AND time < now() - 19h30m

I get (without the comment I've added afterwards):

name: test0
time                 Temp_cuve
----                 ---------
2017-02-08T18:36:09Z 38.56
2017-02-08T18:36:28Z 38.56
2017-02-08T18:36:48Z 38.56
2017-02-08T18:37:07Z 38.56
2017-02-08T18:37:27Z 38.56
2017-02-08T18:37:46Z 38.5
2017-02-08T18:38:06Z 38.5
2017-02-08T18:38:25Z 38.56
2017-02-08T18:38:44Z 38.5
2017-02-08T18:39:06Z 38.5 # --> Last value plotted in Grafana
2017-02-08T18:39:25Z 38.56
2017-02-08T18:39:45Z 38.56
2017-02-08T18:40:04Z 38.5
2017-02-08T18:40:24Z 38.56
2017-02-08T18:40:43Z 38.5
2017-02-08T18:41:03Z 38.5
2017-02-08T18:41:22Z 38.56
2017-02-08T18:41:42Z 38.56
2017-02-08T18:42:01Z 38.5
2017-02-08T18:42:21Z 38.5
2017-02-08T18:42:40Z 38.56
2017-02-08T18:43:00Z 38.5
2017-02-08T18:43:19Z 38.5
2017-02-08T18:43:39Z 38.5
2017-02-08T18:43:58Z 38.5

But in Grafana, which was working well until now, the plotting of all the data has stopped at the value of 19h39m06s as shown in the image below, whereas the query is the same as the previous one used in the CLI :

Grafana screenshot

Do you have any idea to explain this behavior and plot the rest of the data ? Is there for example any point number limit in Grafana ? (the problem stand for all my 10 variables)

Thank you !

Solution

  • Finally find the answer : the problem comes from the new 1.2 version of influxDB where a new feature has not been yet taken into account by grafana. The problem is about a max-row-limit variable to be send as response and the possible chunked data retourned, as discussed on the grafana github.

    To avoid this, instead of modifying the influxDB parameters, I finally just add the $timeFilter variable in my queries to limit the number of row returned by influxDB, like this:

    SELECT Temp_cuve FROM test0 WHERE $timeFilter

    This variable gives the current time range selected for displaying the graphs.