Skip to main content
Understanding these limits helps you plan capacity for data modeling and records in your CDF project. Limits are grouped by area:
  • Data modeling — schema objects, instances, property values, and reserved identifiers
  • Records and streams — stream counts, template limits, and on-request capacity
  • API concurrency — simultaneous instance operations
  • Monitoring — view usage and request higher limits

​
Data modeling

Default limits for schema resources and instances. Your project may have different values depending on configuration.

​
Schema and instance limits

The table lists the default limits for data modeling schema resources and instances.

​
Property value limits

These limits apply to the values of container properties:

​
List size limits

By default, list properties have these size constraints:
  • Direct relations: maximum 100 items
  • Other types: maximum 1,000 items
You can customize the limits by setting the maxListSize parameter on container properties. You can only increase (not decrease) the maxListSize parameter after creating a container property. The maximum values for maxListSize are:
  • 600 for int32 properties with a b-tree index.
  • 300 for int64 properties with a b-tree index.
  • 2,000 for all other property types.

​
Reserved values

The following values are reserved for Cognite internal use and cannot be used in your data models.

​
Views and containers

Reserved externalId values:
  • Boolean, Date, Float32, Float64, Float, Int32, Int64, Int
  • JSONObject, Mutation, Numeric, PageInfo, Query, String
  • Subscription, TimeSeries, File, Sequence, Timestamp
Reserved properties values:
  • createdTime, deletedTime, edge_id, extensions, externalId
  • lastUpdatedTime, node_id, project_id, property_group
  • seq, space, version, tg_table_name, startNode, endNode
Reserved enum value names:
  • "true", "false", "null"

​
Spaces

Reserved space values:
  • cdf, dms, edge, node, pg3, shared, space, mapping, system
  • constraint, edge_source, history, identifier_by_space, index
  • node_source, project, property_group_constraint, property_group_index
  • property_group, property, version_info

​
Records and streams

This section lists resource limits for records and streams. For an introduction to the concepts, see About records and streams. For sizing guidance and how the limits shape your design, see Choosing and sizing streams.

​
Stream limits per project

The table lists the default stream limits for a CDF project. The limits can be different for your CDF project depending on your specific configuration. You can mix streams from different templates within the total active stream cap. To raise these limits, see Request additional capacity.

​
Stream template limits

Stream templates are in betaStream templates aren’t in production yet. We may add templates, adjust limit values, or expose more options when you create a stream. Any changes apply only to new streams. Streams you’ve already created keep the limits that were in effect at creation time.For capacity planning and template selection, see Choosing and sizing streams.
The stream template you select fixes the stream’s limits at creation time. You can’t change a stream’s template or limits after you create it. The table lists the default limits for each stream template. For what each template is designed for and how to select one, see Stream templates and Choose a default template.

​
On-request capacity

For enterprise customers with larger workloads, Cognite can enable higher limits on request. This is not automatic with your subscription. For the request process, see Request additional capacity. For solution architecture estimates, Cognite has provisioned: Custom data retention periods and additional active streams are also available on request. Routine requests typically take 1 to 2 weeks. If your workload exceeds these figures, contact your Cognite representative to discuss your needs on a case-by-case basis. No API is available today to list which stream templates are enabled for your project. Contact your Cognite representative to confirm what’s available beyond the defaults. Rate and concurrency limits for the Records and Streams API endpoints are documented in API rate limits.

​
API concurrency limits

Concurrency limits control how many data modeling instance API operations can run simultaneously on your CDF project. The table lists the default concurrency limits for a CDF project. The limits can be different for your CDF project depending on your specific configuration. Transformations can use a maximum of 75% of the budget while the remaining 25% is reserved for other clients. Concurrency limits are applied project-wide, not per client. If a request exceeds any of the concurrency limits, it will be throttled with a 429 Too Many Requests response. Learn more about API rate limits.
If you are using official Cognite SDKs, throttled requests are automatically retried with exponential backoff. If you are making a request without using an official SDK, make sure that you are using an appropriate retry strategy to handle 429 responses.

​
Monitoring resource usage

To view the current limits and usage statistics for your CDF project:
1

Navigate to the Data models page

In the Data fusion workspace, select Data models.
2

View the limits and usage statistics

On the Data models page, under Storage, select See all to view your project-specific limits and current usage statistics.

​
Retrieve limits programmatically

To retrieve limits and usage statistics programmatically:

​
Need higher limits?

If you need increased limits for data modeling resources or instances, contact Cognite Support. For records and streams capacity above the defaults, see On-request capacity or Request additional capacity.
Last modified on August 17, 2026