Skip to main content

What's new in ArcGIS Velocity

The November 2026 release includes quality improvements, enhancements, and new functionality throughout ArcGIS Velocity. Highlights are provided below.

For details, refer to the What's new in ArcGIS Velocity ArcGIS Blog article.

General

General updates include the following:

  • Numerous accessibility enhancements were made.

  • Back-end libraries and dependencies were updated.

  • You can filter the list of feeds, data sources, and output types.

  • You can view a reference table with tags for each feed, source, and output type.

  • Partner feeds display partner icons and partner levels.

  • The ArcGIS Arcade AI Assistant is available in the Arcade editor in Velocity. The Arcade editor also includes a debugger.

Feeds and sources

The following are enhancements to feeds and sources in Velocity:

  • Velocity supports the following new feed types:

  • Added a new parameter, Get latest locations only, in the Geotab (LogRecord) feed. This parameter returns only the latest location for each device.

  • The minimum polling interval of the HTTP Poller feed is 5 seconds. Previously, it was 10 seconds.

  • Velocity feeds support geometry with M values or linear referencing measures from linear referencing system (LRS) sources. While configuring a feed, in the Identify key fields step, under the Location section, you can set the M (measure) parameter, which opens the Measuring section where you can specify the Route ID Field, Measure Field, To Route ID Field, and To Measure Field.

  • You can reload source data on a schedule without restarting the analytic.

  • The MQTT feed supports certificate-based authentication. You can use certificates to connect to brokers, such as Azure and AWS, that require them.

  • In the gRPC feed, the Flatten and Flatten arrays options are available when the data format is JSON or XML.

  • The HTTP Simulator feed supports basic authentication. Optionally, you can provide a username and password in the user interface.

  • Feeds now reuse one access token instead of requesting a new token for each request.

  • The Azure Event Hub feed shows a warning in the logs when another consumer that uses the same consumer group takes over the receiver and disconnects the feed. Previously, the disconnect was not shown as a warning.

  • Feature Layer inputs do not fail at startup or preview when token refresh and retries overlap with ArcGIS token refresh handling. Previously, scaling completed successfully, but retries logged the NullPointerException error message because token timer was stopped or cleared, and then canceled by another retry or refresh.

  • A running feed picks up a new refresh token after the scheduled token renewal replaces it. Previously, the feed could keep using the old token, which ArcGIS Online had invalidated, and fail with a 498 invalidated refresh_token error on each run until its own token reload timer ran.

  • When configuring a Stream Layer feed, data sampling does not fail when a WHERE clause is applied to an esriFieldTypeBigInteger (64-bit Integer) field. Previously, an error message was logged.

  • The HTTP Poller feed supports repeated URL parameters. The feed sends every value of a repeated query parameter, in the order you entered them. Previously, only the last value was sent and the others were dropped.

Outputs

The following are enhancements to outputs in Velocity:

  • Velocity supports the following new output type:

  • In the Microsoft Teams output, the Microsoft Teams webhook URL parameter's property has been changed to password because the URL contains a signature.

  • The following enhancements have been made to the Feature Layer (new) or Feature Layer (existing) outputs:

    • You can connect more than one pipeline when the schemas match or one is a subset of the other. Fields that a pipeline does not send are not overwritten.

    • Added a new parameter, Keep the newer feature for a Track ID based on Start Time. You can keep the newer feature for a track ID based on its start time instead of its arrival time. You set this in Advanced options on feature layer outputs.

    • Added a new parameter, Simplify feature geometry, to the feature layer outputs and the Stream Layer output. This parameter simplifies feature geometries to ensure they are written to the database correctly.

  • The output node for a new spatiotemporal feature layer shows that the layer is stored and served in WGS84 (4326). Previously, the output node for a new spatiotemporal feature layer showed the spatial reference of the upstream data, even though the layer is always stored and served in WGS84 (4326).

  • The RabbitMQ output metrics show the correct feature rate while messages are published. Previously, the metrics showed 0 f/s even when the output was publishing messages successfully.

  • Velocity returns a validation error when you create a Stream Layer output with the same name as an existing stream layer, feed, or spatiotemporal feature layer. Previously, you could create a stream layer with the same name as a feed, even though stream layer names must be unique in the organization.

Analytics

The following are enhancements to analytics of Velocity:

  • Velocity supports the following new tools:

  • You can add a new field to an existing Velocity-hosted feature layer directly from the Layers page without recreating the feature service.

  • Added a new parameter, Prevent wrong-way snapping on one-way roads, in the Snap to Network tool. This parameter specifies whether a point feature is snapped to a one-way road segment when it is traveling opposite to the road's permitted driving direction.

  • Updated the icon of the Multi-Merge tool in the user interface.

  • Added a parameter, Target time window, in the Detect Gaps tool. This parameter anticipates the longest time period between observations for any single track.

  • A warning appears when you select two or more fields for the Track ID fields parameter for the following tools: Detect Gaps, Detect Incidents, Control Event Volume, Filter by Geometry, Find Dwell Locations, Join Features, Reconstruct Tracks, Snap to Network, Calculate Journeys, and Calculate Motion Statistics.

  • You can write text from a smaller string field to a larger string field. Previously, for example, if the input field is defined as a text field with 250 characters and the output field contains 2048 characters, the output field would be empty.

  • Standard logs now show a message when a real-time analytic recovers after its Kafka streams internal topics are missing and its processing state is lost. Previously, this message appeared only in Advanced logs.

  • The Failed to fetch Tenant Setting From ConfigStore warning no longer appears in the user interface in Advanced logs. Velocity uses default tenant settings without logging a warning when a tenant has no saved settings.

  • Analytics reuse one access token instead of requesting a new token for each request.

  • Velocity now recreates a Kafka topic that belongs to a real-time analytic if the topic is deleted, so the analytic can recover without a manual restart. Previously, only feed topics were recreated. If an analytic's topic was deleted, it logged missing topic errors until you stopped and restarted it.

  • Big data analytics handle 504 Gateway Timeout responses when Velocity retrieves the feature layer or stream layer service definition for a source or output during a run. Previously, a temporary gateway error could cause the run to fail with the message Velocity was unable to retrieve the feature service definition.

  • You can type / in tool settings in the analytic designer. Previously, pressing / moved the pointer to the Add Node search box instead of entering the character. The / shortcut for Add Node is removed.

  • The Configure Stream Layer window keeps your selected filter, such as My Content, when you click Previous step. Previously, the filter changed to By URL and filled in the service URL, and selecting Next returned an error that the URL was not accessible.

  • The Detect Incidents tool with an enter or exit relationship and the Close on Expiration parameter turned on closes expired incidents without errors. Previously, the ArrayIndexOutOfBoundsException process error was logged after an incident reached its expiration.

  • The Map Fields tool does not log the warning Map Fields tool geometry calculation expression failed for 0 records when no geometries are null. Previously, analytics, such as the Winter Operations Analytics big data analytic, logged this warning even though no records failed.

  • The Add XYZ Values tool does not add two fields named z to the output schema when you name an output field z and the input schema has no z field. Previously, the schema had duplicate z fields and the tool node showed a validation error.

  • The Target time window parameter is honored in the Join Features, Filter by Geometry, and Detect Incidents tools. Previously, observations stayed in memory longer than the Target time window, so an enter or exit incident could still be detected against an earlier observation that was older than the window.

  • Big data analytics do not intermittently log the error Error reporting logs for component in the Logs tab.