You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I have recently started an internship at a section of the Max Plank institute.
Their current database system does not have good options to plot and monitor multiple streams of data, however their database has a JSON API that is almost compatible with your Grafana JSON Datasource.
As such I was hoping to make use of your adapter as well as Grafana to get a better overview of their measurements.
I am able to properly fetch data from an endpoint and extract the necessary timestamps and data values.
However, they are using a timestamp with nanosecond precision, given in a numeric format (as integer).
This value is parsed just fine, but Grafana itself refuses to accept this data column as timestamp, which does not allow it to be plotted in any of the time-series based panels.
I believe the easiest way to make this work is to add another column type, similar to "time", perhaps called "time as ns". It would
merely have to divide the given timestamp by 1000000 and generate a timestamp that Grafana can work with.
The loss of precision should not be an issue - the data example above already shows that some of their measurement streams don't make use of the ns precision.
I would be willing to dig in to the code and add this feature myself, however I have little experience in writing Grafana plugins, and was wondering how hard this would be to implement with your current system architecture.
Perhaps it just requires a small addition to the initial data parser?
Best regards,
D. Bailey
The text was updated successfully, but these errors were encountered:
Greetings Mr. Olsson,
I have recently started an internship at a section of the Max Plank institute.
Their current database system does not have good options to plot and monitor multiple streams of data, however their database has a JSON API that is almost compatible with your Grafana JSON Datasource.
As such I was hoping to make use of your adapter as well as Grafana to get a better overview of their measurements.
I am able to properly fetch data from an endpoint and extract the necessary timestamps and data values.
However, they are using a timestamp with nanosecond precision, given in a numeric format (as integer).
This value is parsed just fine, but Grafana itself refuses to accept this data column as timestamp, which does not allow it to be plotted in any of the time-series based panels.
An example API response would be:
I believe the easiest way to make this work is to add another column type, similar to "time", perhaps called "time as ns". It would
merely have to divide the given timestamp by 1000000 and generate a timestamp that Grafana can work with.
The loss of precision should not be an issue - the data example above already shows that some of their measurement streams don't make use of the ns precision.
I would be willing to dig in to the code and add this feature myself, however I have little experience in writing Grafana plugins, and was wondering how hard this would be to implement with your current system architecture.
Perhaps it just requires a small addition to the initial data parser?
Best regards,
D. Bailey
The text was updated successfully, but these errors were encountered: