When the DataNodeServer is first started it runs get_fields (done in onJoin). However, for some servers the fields list is unpopulated until data has been collected. This is true currently for the live monitoring thermometry_server.py.
The bug that occurs as a result of this is the user expects to see live data in Grafana, but instead no data is returned. The conditions for this to occur are:
the sisock stack is running
the OCS Agents which collect the data are running
the user has an already configured Grafana dashboard/panel that should display the live data.
When data collection starts the field list is not populated with known fields (because get_fields only runs at startup), the server returns no data.
The user can cause get_fields to be re-run by editing a panel in Grafana and clicking on the drop down box in which they select the field to plot, after which the data will begin to show up.
We should consider how to populate the fields list to avoid this bug.
When the
DataNodeServer
is first started it runsget_fields
(done inonJoin
). However, for some servers the fields list is unpopulated until data has been collected. This is true currently for the live monitoringthermometry_server.py
.The bug that occurs as a result of this is the user expects to see live data in Grafana, but instead no data is returned. The conditions for this to occur are:
When data collection starts the field list is not populated with known fields (because
get_fields
only runs at startup), the server returns no data.The user can cause
get_fields
to be re-run by editing a panel in Grafana and clicking on the drop down box in which they select the field to plot, after which the data will begin to show up.We should consider how to populate the fields list to avoid this bug.