Skip to main content

Resource capacity

ArcGIS Velocity allows you to create, share, and manage real-time ingestion and analysis tasks, which are processing (computing) tasks. Analytics can also write feature records as output to feature layers, which are hosted in ArcGIS Enterprise and stored in a relational data store or a spatiotemporal big data store. All of these features have a compute, memory, and storage cost.

Compute capacity

Each running feed or analytic consumes compute resources, and this consumption varies by data characteristics, velocity, and volume. Consider the following examples:

  • Feeds ingesting tens of thousands of observations per second consume more compute capacity than feeds that support lower velocities.

  • Real-time analytics processing higher velocity data consume more compute capacity than analytics processing lower velocity data streams.

Administrators can view the percentage of compute used on the Home page under the Resource Utilization section in the Velocity app.

Memory capacity

Memory is used by Velocity processes, such as feeds and analytics running on the machine, to handle observation data and analytic results. Memory capacity is relative to the total memory allocated to each process running Velocity and does not account for other processes that use memory.

Administrators can view the percentage of memory being used on the Home page under the Resource Utilization section in the Velocity app. Similar to compute capacity, memory usage is divided among feeds, real-time analytics, stream processes, system operations (control plane or metrics), and the remaining available capacity of the system.

Note:

Administrators can adjust the amount of memory allocated to each service that makes up the Velocity site by using the administrative API.

Running items

Administrators can view the number of running items on the compute and memory pages in the Velocity app. Both pages report the list of feeds and real-time analytics that are actively running. Administrators can stop running items directly from these pages.