Practitioner ⭐️

Solved

Hosted HiveMQ Extractor defaults to OLD Time Series and not CogniteTimeSeries

Forum|Forum|10 months ago
September 25, 2025
5 replies
86 views

Is there a way to configure a Hosted HiveMQ Extractor to extract the data into CogniteTimeSeries and not the traditional Time Series?

Best answer by Einar Omang

Ah, you’re using the built-in sparkplug format? The frontend doesn’t support this yet, so you will have to use the API, but we do support setting data modelling space on outputs of the sparkplug mapping, see https://api-docs.cognite.com/20230101/tag/Jobs/operation/create_jobs

I’m realizing the docs there are actually missing sparkplug, but the config is the same as the Cognite format type, i.e.

"format": {
    "type": "sparkplug",
    "dataModels": {
        "space": "my-space"
    }
}

Adding this to the frontend is planned, I believe.

E

Einar Omang

Practitioner ⭐️⭐️⭐️

I’m assuming you refer to the hosted MQTT extractor. We do not have a hosted extractor for HiveMQ in particular.

Yes, see https://docs.cognite.com/cdf/integration/guides/extraction/hosted_extractors/hosted_extractors_custom_mappings#data-points

Simply add `space` to the datapoint output. It will automatically create CogniteTimeSeries if they do not already exist.

For creating metadata you need to use the `node` target, and manually target the `CogniteTimeSeries` view (or some view implementing that).

G

Gerardo ChapaAuthor

Practitioner ⭐️

Thanks @Einar Omang, that answers my question. However does it need to be a custom message to have the space configured, couldn’t it be configured on the extraction pipeline? The MQTT broker is not setup only to serve CDF, and we cannot change the message being sent.

E

Einar Omang

Practitioner ⭐️⭐️⭐️

I would advise you to look into how to use custom formats in the hosted extractors, you can apply arbitrary transformations to the incoming messages. I would be surprised if you didn’t already have this, as it is highly unlikely that the format of the messages from the MQTT broker matches the expected format for CDF.

See https://docs.cognite.com/cdf/integration/guides/extraction/hosted_extractors/hosted_extractors_custom_mappings

G

Gerardo ChapaAuthor

Practitioner ⭐️

Right, the MQTT broker is publishing in Sparkplug B format, would it be possible to map that and create a custom mapping?

E

Einar OmangAnswer

Practitioner ⭐️⭐️⭐️

Ah, you’re using the built-in sparkplug format? The frontend doesn’t support this yet, so you will have to use the API, but we do support setting data modelling space on outputs of the sparkplug mapping, see https://api-docs.cognite.com/20230101/tag/Jobs/operation/create_jobs

I’m realizing the docs there are actually missing sparkplug, but the config is the same as the Cognite format type, i.e.

"format": {
    "type": "sparkplug",
    "dataModels": {
        "space": "my-space"
    }
}

Adding this to the frontend is planned, I believe.

Sign up

Welcome to Cognite Hub