Reference · built on requestConnectorPLC, SCADA & MES

Cognite Data Fusion

Assets, time series, data points, events, files, and data-model instances from Cognite Data Fusion through the CDF API.

Request accessPricing ↗

About

Cognite Data Fusion is an industrial data platform. Its API lives at https://{cluster}.cognitedata.com/api/v1/projects/{project} and holds assets in hierarchies, time series with numeric or string data points, events with start and end times, files, sequences, RAW tables, and data-model instances. Extraction pipelines record which extractors feed the project and whether their runs succeeded.

Through Fibric, an operator lists assets and events, follows data point subscriptions for the time series it watches, and proposes what follows: a new event on an asset, an updated instance, or a data point on a calculated series. You approve. Each write is sent once and leaves a receipt: what changed, why, and how to undo it.

This is a reference listing. It documents what Fibric would read from Cognite Data Fusion and what it could propose, based on the vendor's published interfaces. Fibric builds it under a managed deployment when you request it; selecting it here installs nothing.

Inputs

  • Assets by hierarchy, filter, search, or aggregate through the Assets endpoints
  • Time series and their data points, latest values, and aggregates through Retrieve data points and Retrieve latest
  • Changes to time series in ingestion order through Data point subscriptions and its list subscription data endpoint
  • Events with start time, end time, description, and metadata through Filter events and Search events
  • Files, sequences, relationships, labels, and data sets
  • Data-model instances through the Instances query and sync endpoints
  • Extraction pipeline runs with success, failure, and seen statuses

Proposed actions

  • Target capability: propose creating or updating an event on an asset through Create events and Update events
  • Target capability: propose inserting data points on a time series through the insert data points endpoint
  • Target capability: propose creating or updating data-model nodes and edges through the Instances apply endpoint
  • Target capability: propose starting a transformation run through Run transformation

Proposed actions are target capabilities. Every action runs propose-first and needs a validated deployment and the appropriate permissions.

What you can build

  • Explain a pump's alarm from its events and trends

    Alarm events on an asset are read with the time series under it. The operator lays the data points against the event window and drafts a cause summary for the reliability engineer.

    With Root Cause, Alarm Flood

  • Keep a maintenance backlog in sync with CDF events

    Predicted failures become proposed events on the affected asset, with description and metadata filled in. Nothing is created until you approve, and each event carries the evidence used.

    With Predictive Maintenance, Inspection Due

  • Notice when an extractor goes quiet

    Extraction pipeline runs carry seen heartbeats and failure statuses. The operator lists pipelines whose last seen is late and proposes a notice to the data owner.

    With Data Freshness, Integration Health

Requirements

  • A CDF project, its cluster hostname, and an application registered with the project's identity provider
  • Group membership granting capabilities such as timeSeriesAcl:READ for the resources you read or write
  • A cdf-version header such as 20230101 when you pin an API version; -beta and -alpha suffixes mark preview features
  • Data point subscriptions created for the time series you want to follow
Authentication
An OpenID Connect access token from the identity provider configured on the CDF project, sent as Authorization: Bearer; client credentials and authorization code flows are both accepted.

Limits

  • Requests beyond the project's rate or concurrency budget return HTTP 429; Cognite recommends truncated exponential backoff
  • List, Search, and Filter endpoints have a smaller request budget than create, retrieve, update, and delete
  • A subscription holds at most 10000 time series, a time series joins at most 10 subscriptions, and updates are kept 7 days
  • Parallel retrieval uses at most 10 partitions; some resources enforce that cap

Access and pricing

Reference listing. Fibric builds the connector under a managed deployment when you request it. Your quote covers the build, capabilities, usage, and support.

Request Cognite Data Fusion ↗

Questions and answers

How does Fibric follow time series changes without re-reading history?
With data point subscriptions. A subscription lists the time series to follow, explicitly or by filter. The list subscription data endpoint returns upserts and deleted ranges in ingestion order with a cursor. Updates are retrievable for 7 days after they are written.
Which authentication does the CDF API accept?
OpenID Connect tokens issued by the project's identity provider, passed as Authorization: Bearer. Client credentials and authorization code flows are both listed. Service accounts, API keys, and the /login endpoint were removed.
How is the API versioned?
By calendar date. Pass cdf-version: 20230101 to pin a version. Beta features use a -beta suffix and alpha features an -alpha suffix on the same header.
Ask about Cognite Data Fusion

Ask about the capabilities and requirements in this listing.

For project-specific requirements, contact Fibric.

This connector listing is developed, published, and supported by Fibric. It describes integration with Cognite Data Fusion through published interfaces. Third-party names and logos identify the systems an integration connects to; they are the property of their respective owners, who are not affiliated with Fibric and do not sponsor or endorse this listing. Trademark policy