Scalability#

Observability system is designed to scale both horizontally and vertically to handle a large number of Edge Nodes. To achieve balance between performance of multiple components at scale, the system requires careful tuning.

Edge Orchestrator ships with a set of pre-defined profiles that correspond to different number of expected Edge Nodes to be managed, so it’s mostly relevant for Edge Node Observability Stack. The profiles tune both the horizontal and vertical scaling of individual components of the system to assure optimal performance of the stack.

The available predefined profiles, which configure observability components via Argo* CD app overrides, are:

  • Development

    • o11y-dev.yaml - default development profile consuming minimal resources.

  • Production (cloud-based)

    • o11y-release.yaml - default profile for up to 100 Edge Nodes.

    • o11y-release-large.yaml - profile for 1000 Edge Nodes (and more).

  • Production (on-premises)

    • o11y-onprem.yaml - default profile for up to 100 Edge Nodes.

    • o11y-onprem-1k.yaml - profile for up to 1,000 Edge Nodes.

For more information on the system requirements regarding scalability, refer the following guides: