Prometheus component not exporting metrics for individual entities using dot notation

I had prometheus and grafana configured some 12 months ago and to pull data from Home Assistant and display it in Grafana dashboards.

I have now reconfigured this using this docker stack but the query format seems to have changed as all my Grafana dashboards are broken and incompatible.

It is the most puzzling thing:

Previously, all i had to do was enter the Home Assistant entity_id in the Grafana query text field and a graph would be produced. This is because the prometheus component would export each sensor entity id as a distinct metrics for Prometheus

For Example:


Now, this simplified querying syntax is gone it appears. The example expression below is now equired to achieve the same effect.



I need to know the unit of the metric I wish to graph (“hass_sensor_unit_c”) and then need to check in the query inspector in Grafana which metrics are available for that unit.

Is there something i am missing? Is there a way to configure Prometheus to scrape data like it used to? I do not want this grouping by metric unit, instead i want seperate ones like before.

I already posted this on the Prometheus forum and they suggest it has to do with Home Assistant and how the Prometheus component formats the information.
The compoent documentation does not show any configuration options for exporting individual entities as metrics. This used to work, wondering what happened?

I’m still confused how this changed and why I am the only one facing this breaking change. The prometheus component is grouping entities by unit and it no longer using dot notation.

Hi Danny,

I think I have hit exactly this issue too.

I was able to find what i needed by using the following PROMQL query to list off all the metrics from my hass job

sum by(__name__)({job="hass"})

And then to narrow them down like this


how would you use this to create a graph?