Request for Guidance on Efficient Data Visualization with Grafana and Superset using TimescaleDB about stellio-context-broker

[bobeal]

Hi Michael,

Sorry for late reply.

Actually, we are using a lot Grafana and Superset in our typical deployments.

When using Grafana, we get data directly from the DB (as Grafana is a real-time oriented tool, this is by far the most efficient way to work with). And we are doing using such query:

SELECT
  time AS "time",
  measured_value AS "temperature",
  temporal_entity_attribute.entity_id as entity_id
FROM attribute_instance
LEFT JOIN temporal_entity_attribute ON temporal_entity_attribute.id = attribute_instance.temporal_entity_attribute
WHERE
  $__timeFilter(time) AND
  temporal_entity_attribute.attribute_name = 'https://vocab.egm.io/temperature' AND
  temporal_entity_attribute.entity_id in ($Room) -- where Room is a dashboard variable
ORDER BY 1

For Superset, this is getting a bit more complex since we typically want to apply a denormalization process on the data. To do this, we use Apache NiFi (https://nifi.apache.org): using the NGSI-LD Subscription API, a notification is sent to a flow in NiFi when an entity of interest is updated, then the flow retrieves all the entities that the entity is in relation with (its relationships) and export the result in a "big" table which is easier to work with a BI oriented tool like Superset.

The specific NiFi processor we are using to do this is in https://github.com/easy-global-market/nifi-ngsild-postgresql. It is unfortunately missing some documentation right now but it can be an opportunity to at least add some instructions on how to set it up and use it.

Let me know.

Benoit.